Computer Science and Engineering Theory and

Applications Coll. download

https://textbookfull.com/product/computer-science-and-
engineering-theory-and-applications-coll/

Download more ebook from https://textbookfull.com

We believe these products will be a great fit for you. Click
the link to download now, or visit textbookfull.com
to discover even more!

Advances and Applications in Computer Science,

Electronics and Industrial Engineering: Proceedings of
the Conference on Computer Science, Electronics and
Industrial Engineering (CSEI 2019) Jyrki Nummenmaa
https://textbookfull.com/product/advances-and-applications-in-
computer-science-electronics-and-industrial-engineering-
proceedings-of-the-conference-on-computer-science-electronics-
and-industrial-engineering-csei-2019-jyrki-numme/

Algebra Topology Differential Calculus and Optimization

Theory For Computer Science and Engineering Jean
Gallier

https://textbookfull.com/product/algebra-topology-differential-
calculus-and-optimization-theory-for-computer-science-and-
engineering-jean-gallier/

Cognitive Informatics, Computer Modelling, and

Cognitive Science: Theory, Case Studies, and
Applications: Volume 1: Theory, Case Studies, and
Applications 1st Edition Ganesh R. Sinha (Editor)
https://textbookfull.com/product/cognitive-informatics-computer-
modelling-and-cognitive-science-theory-case-studies-and-
applications-volume-1-theory-case-studies-and-applications-1st-
edition-ganesh-r-sinha-editor/

Computer Science Theory and Applications 15th

International Computer Science Symposium in Russia CSR
2020 Yekaterinburg Russia June 29 July 3 2020
Proceedings Henning Fernau
https://textbookfull.com/product/computer-science-theory-and-
applications-15th-international-computer-science-symposium-in-
russia-csr-2020-yekaterinburg-russia-
Transactions on Engineering Technologies World Congress
on Engineering and Computer Science 2015 Amouzegar

https://textbookfull.com/product/transactions-on-engineering-
technologies-world-congress-on-engineering-and-computer-
science-2015-amouzegar/

Probability and statistics with reliability queueing

and computer science applications Trivedi

https://textbookfull.com/product/probability-and-statistics-with-
reliability-queueing-and-computer-science-applications-trivedi/

Advances in Computer Science for Engineering and

Education II Zhengbing Hu

https://textbookfull.com/product/advances-in-computer-science-
for-engineering-and-education-ii-zhengbing-hu/

Advances in Computer Science for Engineering and

Education III Zhengbing Hu

https://textbookfull.com/product/advances-in-computer-science-
for-engineering-and-education-iii-zhengbing-hu/

Academic English for Computer Science An English for

Specific and Academic Purposes Course for International
students of Computer Science Computer Engineering
Information and Communication Systems Rizopoulou
https://textbookfull.com/product/academic-english-for-computer-
science-an-english-for-specific-and-academic-purposes-course-for-
international-students-of-computer-science-computer-engineering-
Studies in Systems, Decision and Control 143

Mauricio A. Sanchez
Leocundo Aguilar
Manuel Castañón-Puga
Antonio Rodríguez-Díaz Editors

Computer
Science and
Engineering—
Theory and
Applications
Studies in Systems, Decision and Control

Volume 143

Series editor
Janusz Kacprzyk, Polish Academy of Sciences, Warsaw, Poland
e-mail: [email protected]
The series “Studies in Systems, Decision and Control” (SSDC) covers both new
developments and advances, as well as the state of the art, in the various areas of
broadly perceived systems, decision making and control- quickly, up to date and
with a high quality. The intent is to cover the theory, applications, and perspectives
on the state of the art and future developments relevant to systems, decision
making, control, complex processes and related areas, as embedded in the fields of
engineering, computer science, physics, economics, social and life sciences, as well
as the paradigms and methodologies behind them. The series contains monographs,
textbooks, lecture notes and edited volumes in systems, decision making and
control spanning the areas of Cyber-Physical Systems, Autonomous Systems,
Sensor Networks, Control Systems, Energy Systems, Automotive Systems,
Biological Systems, Vehicular Networking and Connected Vehicles, Aerospace
Systems, Automation, Manufacturing, Smart Grids, Nonlinear Systems, Power
Systems, Robotics, Social Systems, Economic Systems and other. Of particular
value to both the contributors and the readership are the short publication timeframe
and the world-wide distribution and exposure which enable both a wide and rapid
dissemination of research output.

More information about this series at http://www.springer.com/series/13304

Mauricio A. Sanchez Leocundo Aguilar
•

Manuel Castañón-Puga Antonio Rodríguez-Díaz

•

Editors

Computer Science
and Engineering—Theory
and Applications

123
Editors
Mauricio A. Sanchez Manuel Castañón-Puga
Calzada Universidad Calzada Universidad
Tijuana, Baja California Tijuana, Baja California
Mexico Mexico

Leocundo Aguilar Antonio Rodríguez-Díaz

Calzada Universidad Calzada Universidad
Tijuana, Baja California Tijuana, Baja California
Mexico Mexico

ISSN 2198-4182 ISSN 2198-4190 (electronic)

Studies in Systems, Decision and Control
ISBN 978-3-319-74059-1 ISBN 978-3-319-74060-7 (eBook)
https://doi.org/10.1007/978-3-319-74060-7
Library of Congress Control Number: 2017963287

© Springer International Publishing AG, part of Springer Nature 2018

This work is subject to copyright. All rights are reserved by the Publisher, whether the whole or part
of the material is concerned, specifically the rights of translation, reprinting, reuse of illustrations,
recitation, broadcasting, reproduction on microfilms or in any other physical way, and transmission
or information storage and retrieval, electronic adaptation, computer software, or by similar or dissimilar
methodology now known or hereafter developed.
The use of general descriptive names, registered names, trademarks, service marks, etc. in this
publication does not imply, even in the absence of a specific statement, that such names are exempt from
the relevant protective laws and regulations and therefore free for general use.
The publisher, the authors and the editors are safe to assume that the advice and information in this
book are believed to be true and accurate at the date of publication. Neither the publisher nor the
authors or the editors give a warranty, express or implied, with respect to the material contained herein or
for any errors or omissions that may have been made. The publisher remains neutral with regard to
jurisdictional claims in published maps and institutional affiliations.

Printed on acid-free paper

This Springer imprint is published by the registered company Springer International Publishing AG part
of Springer Nature
The registered company address is: Gewerbestrasse 11, 6330 Cham, Switzerland
Preface

The fields of computer science and computer engineering are rich within them-
selves such that they are comprised of a vast amount of areas, such that most books
decide to better separate their contents into specialized areas, ignoring the rest
of their field. But this book takes on a different point of view when compared to
most of existing literature, that is, to collect from different areas within these two
fields to demonstrate the great variety which exists, where by harboring different
contributed research chapters in a single book, instead of granular specialized areas,
the concept of the complete fields is once again grasped.
As computer science and computer engineering are two different fields, they
have more in common than differences; therefore, the objective of this book is to
convey their involvement in society by showing advances in theoretical practices,
new uses for existing concepts, and applications.
This book is intended as a reference for pre-graduate students who wish to know
more about advances in the fields of computer science and computer engineering, or
acquire ideas as to what types of areas can be researched, exploited, or applied; for
graduate students and researchers, who might need to know the current state of
included contributed research, as well as acquire ideas for their own research; and
for professionals, who would want to know the current state of their fields, get ideas
to solve problems at their workplace, or simply to get a sense of direction as to
where these fields are going.
The contents of this book traverse various areas in the fields of computer
science and computer engineering, such as software engineering, computational
intelligence, artificial intelligence, complex systems, systems, engineering, and
embedded systems.

v
vi Preface

Provided diversity in the contents of this book is the strength between its covers,
we hope that readers enjoy our book and find some utility within its covers. And
that they start seeing that the fields of computer science and computer engineering
can also be seen as complete and diverse fields, instead of only parts of them.

Tijuana, Mexico Mauricio A. Sanchez

November 2017 Leocundo Aguilar
Manuel Castañón-Puga
Antonio Rodríguez-Díaz
Contents

A Comprehensive Context-Aware Recommender

System Framework . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1
Sergio Inzunza and Reyes Juárez-Ramírez
Executive Functions and Their Relationship with
Interaction Design . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 25
Andrés Mejía Figueroa, J. Reyes Juárez Ramírez and David Saldaña Sage
Integrating Learning Styles in an Adaptive Hypermedia System with
Adaptive Resources . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 49
Carlos Hurtado, Guillermo Licea and Mario Garcia-Valdez
On Modeling Tacit Knowledge for Intelligent Systems . . . . . . . . . . . . . . 69
Violeta Ocegueda-Miramontes, Antonio Rodríguez-Díaz, Juan R. Castro,
Mauricio A. Sanchez and Olivia Mendoza
Influence of the Betweenness Centrality to Characterize the Behavior
of Communication in a Group . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 89
K. Raya-Díaz, C. Gaxiola-Pacheco, Manuel Castañón-Puga, L. E. Palafox
and R. Rosales Cisneros
Multi-layered Network Modeled with MAS and Network Theory . . . . . 103
Jose Parra, Carelia Gaxiola and Manuel Castañón-Puga
A Fuzzy Inference System and Data Mining Toolkit for Agent-Based
Simulation in NetLogo . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 127
Josue-Miguel Flores-Parra, Manuel Castañón-Puga,
Carelia Gaxiola-Pacheco, Luis-Enrique Palafox-Maestre, Ricardo Rosales
and Alfredo Tirado-Ramos
An Approach to Fuzzy Inference System Based Fuzzy
Cognitive Maps . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 151
Itzel Barriba, Antonio Rodríguez-Díaz, Juan R. Castro
and Mauricio A. Sanchez

vii
viii Contents

Detecting Epilepsy in EEG Signals Using Time, Frequency

and Time-Frequency Domain Features . . . . . . . . . . . . . . . . . . . . . . . . . . 167
D. E. Hernández, L. Trujillo, E. Z-Flores, O. M. Villanueva
and O. Romo-Fewell
Big Data and Computational Intelligence: Background, Trends,
Challenges, and Opportunities . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 183
Sukey Nakasima-López, Mauricio A. Sanchez and Juan R. Castro
Design of a Low-Cost Test Plan for Low-Cost
MEMS Accelerometers . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 197
Jesús A. García López and Leocundo Aguilar
Evaluation of Scheduling Algorithms for 5G Mobile Systems . . . . . . . . 213
Christian F. Müller, Guillermo Galaviz, Ángel G. Andrade, Irina Kaiser
and Wolfgang Fengler
User Location Forecasting Based on Collective Preferences . . . . . . . . . . 235
Jorge Alvarez-Lozano, J. Antonio García-Macías and Edgar Chávez
Unimodular Sequences with Low Complementary
Autocorrelation Properties . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 259
Israel Alejandro Arriaga-Trejo and Aldo Gustavo Orozco-Lugo
A Comprehensive Context-Aware
Recommender System Framework

Sergio Inzunza and Reyes Juárez-Ramírez

Abstract Context-Aware Recommender System research has realized that effec-

tive recommendations go beyond recommendation accuracy, thus research has paid
more attention to human and context factors, as an opportunity to increase user
satisfaction. Despite the strong tie between recommendation algorithms and the
human and context data that feed them, both elements have been treated as sepa-
rated research problems. This document introduces MoRe, a comprehensive soft-
ware framework to build context-aware recommender systems. MoRe provides
developers a set of state of the art recommendation algorithms for contextual and
traditional recommendations covering the main recommendation techniques exist-
ing in the literature. MoRe also provides developers a generic data model structure
that supports an extensive range of human, context and items factors that is
designed and implemented following the object-oriented paradigm. MoRe saves
developers the tasks of implementing recommendation algorithms, and creating a
structure to support the information the system will require, proving concrete
functionality, and at the same time is generic enough to allow developers adapt its
features to fit specific project needs.

1 Introduction

Recommender systems or recommendation systems are information filtering tools

that alleviate information overload to users, mainly by predicting the preference of
the user for certain items and presenting the users the items more likely to be
consumed by the user [1].
Context-Aware Recommender Systems (CARS) aims to further improve recom-
mendation accuracy and user satisfaction by taking into account contextual infor-
mation [2]. The inclusion of this new information into the recommendation process
has proved to help increasing prediction accuracy of recommender systems [3].

S. Inzunza (&)
R. Juárez-Ramírez

Universidad Autónoma de Baja California, Tijuana, Baja California, Mexico
e-mail: [email protected]

© Springer International Publishing AG, part of Springer Nature 2018 1

M. A. Sanchez et al. (eds.), Computer Science and Engineering—Theory
and Applications, Studies in Systems, Decision and Control 143,
https://doi.org/10.1007/978-3-319-74060-7_1
2 S. Inzunza and R. Juárez-Ramírez

CARS are based on the idea that similar user in similar context will like similar
items, and that the user preferences for certain item change according to the con-
textual situation [4].
In the beginning, the research in CARS is leaned toward the development of
new, and optimizing the existing algorithms to support contextual information and
to generate better recommendation by improving the accuracy [5].
In recent years, researchers have become more aware of the fact that effective-
ness of recommender systems goes beyond recommendation accuracy. Thus,
research in the context and human factors has gained increased interest [6] as a
potential opportunity to increase the user satisfaction with the recommendation
results.
Despite the strong tie between the recommendation algorithms and the data
about user and context factors, both elements have been treated as separated
research problems. Because of this separation, to implement a CARS nowadays,
software developers, and architects must divide their efforts into 2 topics: (i) The
recommendation system topic to choose and implement the best algorithm for their
CARS domain; (ii) The second topic that also requires a significant amount of effort
is the user modeling. In the first topic, according to [7], the implementation of
recommendation algorithms can be a complicated task, especially for developers
who are not experts in the field [5]. As for the second topic, software architects need
to come up with a data model that is capable of holding all the needed information
for the algorithms to work, if the model designer has no experience with CARS
over what aspects to model for better recommendations, this can result in an
incomplete and overspecialized model [8], that can limit the functionality of the
recommendation algorithms.
A software framework to support the development of advanced, complex con-
textual recommender systems should provide the developer both main elements of a
CARS, implemented algorithms from different recommendation techniques so
developer can select the one that better work for CARS item domains (e.g. songs,
movies, etc.) [9], as well a data model structure generic enough to work with most
CARS domains and give developers the option to adapt it to suit the specific need
of the project. Such a framework is, to the best of our knowledge, currently no
available. Even when some framework and libraries proposals exist to help in the
creation of contextual recommender systems, they are focused on either modeling
the user and context information [10] and do not include recommendation features,
or focus only in the algorithms for recommendation (like Hybreed [5]) and don’t
considered modeling the user and context information and rely on the information
being in a dataset file, which is not practical for a real-world applications.
In this paper, we introduce MoRe (Modeling and Recommendation framework),
a software framework designed to support the creation of CARS by proving
developers a set of state of the art contextual algorithms, and a generic user model
capable of structure and manage the information required by the CARS. We con-
sider MoRe to be a comprehensive framework, as in a single proposal, provides
means to solve both main problems developers encounter when implementing
CARS (i.e. data modeling and recommendation algorithms). MoRe provide
A Comprehensive Context-Aware Recommender System Framework 3

developers object-oriented classes and methods they can use to store, retrieve and
perform recommendation over the user, context and item data.
MoRe aims at serving as a tool for new CARS development, helping developers
in the algorithm implementation by providing ready to work algorithms from main
recommendation techniques for multidimensional data (transformation and adap-
tation). MoRe also helps in the creation of the data architecture by proving an
extensive class architecture capable of organizing, persisting and retrieving data
about the user, context, and items for most popular recommendation domains, like
movie, song, hotels, restaurants, and travel recommendation among others. For the
data structure, uses GUMORS, a General User Model for Recommender System
presented in our previous work [11].
The rest of this document is organized as follows: Sect. 2 describes the back-
ground and related work, Sect. 3 describes the MoRe framework, Sect. 4 present
some evaluation performed, and Sect. 5 present conclusion and future work.

2 Background and Related Work

The implementation of CARS requires modeling of the data and implementation of

algorithms that use such data to generate predictions, as mentioned earlier, both
tasks are treated separately in literature. The same way, in this section we review
and analyze the most relevant proposals related to modeling the context for adaptive
systems and compare them with [11], the data model used in MoRe. Then we
review and analyze most relevant proposals of software libraries and frameworks
created to generate context-based recommendations.
This section concludes with a set of requirements derived from literature that a
software framework intended to help developers in the creation of CARS should
fulfill.

2.1 Context and Context Awareness

The definitions for context varies depending in the domains of applications, in

computer science Schilit et al. [12] described context as a union of three aspects:
“the location of user, the collection of nearby people and objects, as well as the
change to those objects over time”, but such definition is rather broad. Dey and
Abowd [13] provide another definition that is commonly accepted in computer
science areas:
Context is any information that can be used to characterize the situation of an entity. An
entity is a person, place or object that is considered relevant to the interaction between a
user and an application, including the user and the applications themselves.
4 S. Inzunza and R. Juárez-Ramírez

Certainly, based on this definition of context, the user plays an important role,
the context of a user can include any information that describes his situation, like
his location and time, emotional, mental and physiological information, etc. The
amount and type of information to include in the contextualization of the user
depends on the system and the purpose of such information.
A system is considered to be context-aware if it can express aspects of the
environment and subsequently uses such information to adapt its functionality to
the context of use [14]. Therefore, context-awareness refers to the capability of an
application being aware of its physical environment or situation and responding
proactively and intelligently based on such awareness.
Context-aware systems aim at somehow gathering (through sensing, inferring or
directly asking for) human and environmental phenomena for later assisting users to
archive a desirable quality of living standards [15].
From an informational perspective, context provides information that systems
can use to form a contextual space, which virtually represents the situation of the
user [16].
From a software infrastructural perspective, context provides computing devices
with information about its environment as provided by other parts or modules of the
system (sensing module for example). As a consequence, different ‘types’ or ‘di-
mensions’ of context emerge, e.g. physical and computation dimensions.

2.1.1 Context Representation

There are several techniques that will allow a developer to represent the contextual
information inside a computational system, thus to be used in a CARS. Works like
[17–19] present extensive surveys on the difference of each technique. Next, we
describe and discuss the most commonly used representation techniques at
high-level, and present the main advantages and disadvantages of each one.
• Key-Value models. These models use pairs of a key and value to enumerate
attributes (key) and their values to describe the contextual and user information.
These models are the simplest data structure and are easy to manage, especially
when they have a small amount of data. However, key-value modeling is not
scalable and not suitable to represent complex data models [19]. This technique
is best suited to represent and store temporary information, therefore is
increasingly less used in recent contextual and user models.
• Markup scheme models. These models use a hierarchical data structure formed
by markup tags with attributes and content. To represent the user and context
aspect, markup models use a set of symbols and annotations inserted in a text
document that controls the structure, formatting, and relations among annota-
tions [20]. As markup languages do not provide advanced expressive capabil-
ities, reasoning over the data they represent is hard. Further, retrieval,
interoperability, and re-usability of the data over different models can be diffi-
cult, specifically if such models use different markup schemes [17].
A Comprehensive Context-Aware Recommender System Framework 5

• Ontology-based models. Ontologies represent a description of the concepts and

relationships. Ontologies incorporate semantics into XML-based representation
or Resource Description Framework [21]. Ontology-based context models are
fairly common, because of the formal expressiveness and the ability of
ontologies to support reasoning. However, is hard to construct complete
ontologies and avoid the ambiguity in the ontology [18]. Also the information
retrieval can be computationally intensive and time-consuming when the
amount of data is increased [17].
• Graphical models. These models are mainly based on the Unified Modeling
Language (UML) [22] and Entity-Relationship Model (ERM) [23]. UML is a
standardized general-purpose modeling language used in software architecture
description, which can represent the user and context aspects, as well as its
relations. Graphical models are capable of expressing the modeled aspects by
graphical diagrams, and regarding expressive richness, graphical modeling is
better than markup and key-value modeling as it allows relationships to be
captured into the model [17].
• Object-oriented context model. These technique models the data by using
object-oriented techniques, this offers the full power of object orientation like
encapsulation, reusability, and inheritance. In most cases, this model technique
encapsulates the processing of context at the object level and allows instances to
access context and user information by inheritance mechanism [18]. The main
flaw of this technique is the lack of direct model validation, and when complex
models are created, it may not be supported by limited resources hand-held
devices [19]. Nevertheless, as most of the high-level programming languages
support object-oriented concepts, these models can be integrated into
context-aware systems easily. This makes object-oriented modeling to be used
as code base, run-time, manipulation, and storage mechanism for user and
context models.
As there are various model representation techniques, each with their advantages
and disadvantages, a challenged to overcome in our development was to choose the
one that better suit the need to represent a data model for a context-aware rec-
ommendation systems framework. We opted to create a combination of graphical
modeling, which allowed us to create a reach and expressive data model that
contains the intrinsic relationships of the user, context and items information.

2.2 Frameworks for Context-Aware Recommendations

Context-Aware Recommendation Systems has gained a lot of attention in the

personalization and data mining communities, as a result, there exist a lot of lit-
erature focused on different aspects of CARS. Next, we review and analyze CARS
literature that proposed frameworks aimed at supporting developers in the
6 S. Inzunza and R. Juárez-Ramírez

implementation of these type of systems and compares how each of the presented
proposals relates to MoRe.
MyMediaLite [24] is a recommendation library implemented in C# and aimed
towards recommending items based on collaborative filtering recommendation
technique. Even when MyMediaLite was designed to work in traditional (2D)
recommendations, it has been used by [25] as a base-line recommender in
context-aware recommendation. MyMediaLite is similar to our proposal in the fact
that both provide the feature of 2D collaborative-filtering algorithms, but MoRe
goes far beyond providing nD recommendation algorithms.
Hybreed [5] is a Java-based recommendation framework designed specifically to
help developers in the creation CARS. Hybreed focuses on the integration of
external context to recommendation systems and the hybrid combination of dif-
ferent basic recommendation methods. A notable feature of Hybreed is its dynamic
contextualization approach that creates an in-memory temporal model of the user
containing his current situation. Hybreed and MoRe had in common the feature of
proving developers working ready algorithms to generate contextual prediction
based on filtering techniques. But unlike More, Hybreed do not provide recom-
mendation algorithms based on the context-modeling technique. In addition,
Hybreed stands behind in the data modeling technique, as it uses a limited
Key-Value pair, while MoRe uses a context-aware user model.
ConRec [1] is a context-aware recommendation framework focus on serving as a
tool for developers when implementing CARS, ConRec mainly focuses on a
temporal dynamic representation of context, which can automatically aggregate
different contextual features into a single one, so it can be easier for algorithms to
process. ConRec includes its own multi-dimensional recommendation algorithms
that works as an extension to Apache Mahout. This proposal does not describe how
user or context factors are considered into the framework, nor how they are stored.
Compared to MoRe, ConRec includes only one algorithm, that is not supported by
other literature, while MoRe contains implementation for various algorithms that
are well supported by previous research. Also, MoRe present an clear and
well-structured manner of managing the data, while ConRec only mentions that
uses it a user, item, rate and context format, and no information in how to take such
approach to implementation is described.
CoGrec [26] is a theoretical recommendation framework designed to gather
latent features of a group of users based on their social networks then uses such
gathered information to predict items that better suit the preferences of all the
individuals in the group. This proposal is described theoretical, and even when
some evaluation results are presented, no implementations details were given.
CoGrec differs from MoRe, on being a group-based recommendation only, while
MoRe is flexible in the target user(s). Also, MoRe is closer to implementation
providing a working-ready framework.
In [27] a context-aware architecture for mobile recommendation systems are
proposed. The architecture is designed to be generic enough to work in any CARS
domain, and focus on supporting the mobility of CARS, and the communication
between the mobile device and the server where the recommendation take place.
A Comprehensive Context-Aware Recommender System Framework 7

A downfall of this architecture only provides a template for pre-filter recommen-

dations only, while MoRe provides templates and implementation for a range of
recommendation techniques. MoRe is also ahead of this proposal in the manage-
ment and modeling of the data, as the proposal only provides the means of storing
and retrieving the information, while MoRe provides, this functionality along with
the design of a generic data structure.
CARSKit [28] is a Java-based algorithms library, which contains implementa-
tions for various recommendation techniques, for traditional (e.g. average and
collaborative filtering) and contextual (e.g. item splitting and context-aware matrix
factorization) recommendations. Even when CARSKit contains one of the most
complete set of algorithms, these are designed for scientific purposes, focused
mainly in the evaluation of algorithm results, and are not capable of recommending
items to a specific user, which is the main function of the algorithms in a real CARS
implementation. MoRe uses some of the algorithms implementation of CARSKit
and adapted them to be able to recommend items to users. CARSKit and More
differ in their nature, as CARSKit is a set of algorithms, while MoRe is a recom-
mendation framework that as part of its features, contains a set of algorithms.

2.3 Requirements for a Context-Aware Recommender

System Framework

Based on related literature presented above, an in the work of [5, 15, 29, 30], this
section describes a series of high-level functional requirements any software
framework aimed to facilitate the creation of context-aware recommender systems
should fulfill.
First of all, such a framework should include recommendation algorithms from
the most relevant contextual recommendation techniques, which developers can set
up with a few lines of code, and be able to obtain the list of recommended items.
These algorithms should cover the main techniques existing nowadays, which
according to Adomavicius [31] are Contextual Filtering and Contextual Modeling.
To support contextual modeling recommendation, the framework should include
multi-dimensional recommendation algorithms, and to support contextual filtering,
traditional recommendation algorithms are needed as well. Apart from the imple-
mented and ready to work algorithms, the framework should support the integration
of custom techniques and new algorithms.
With regard to the data needed by CARS, the modeling of contextual infor-
mation has been treated separately from the recommendation functions, and none of
the currently existing frameworks support the management of the CARS data in the
same proposal as contextual recommendation algorithms. Therefore, a compre-
hensive contextual recommendation framework should provide a baseline data
structure that supports the data about the users of the system, their context, the items
that the CARS will recommend, and the relations among the different aspects. In
8 S. Inzunza and R. Juárez-Ramírez

this aspect, diversity has a great importance [6], as the framework should support
CARS from different items domains, i.e. the framework should be able to recom-
mend movies, as well as restaurants or pets.
Besides this specific requirements, a framework targeting developers should
meet general software engineering requirements, for example, the set of guidelines
for software architecture proposed by Microsoft Application Architecture Guide
[32]. Such a framework should make easy for developers to adapt existing and to
add new functionality.
According to [5], the expected solution is not as simple as combine all features
from different proposal that meets certain requirements to create a framework that
covers them all. The challenge is to select the most valuable concepts of all these
existing approaches, reduce their programming complexity and combine them, such
that the resulting framework coverts the described requirements with a rich set of
functionalities, and at the same time is easy to use and reduce development effort.

3 The Comprehensive Context-Aware Recommender

System Framework

This section describes the MoRe (Modeling and Recommendation) framework that
was created to support the creation of context-aware recommender systems, by
providing a comprehensive class architecture that can model and manage the
information required for the CARS systems to work, and by incorporating a set of
state of the art recommender algorithms.
MoRe target, but is not limited to, the following users:
• Software developers who want to create a new CARS, either with or without
experienced creating this type of applications. To these users MoRe offers a data
modeling feature that developers can rely on to manage the information needed
by the recommender algorithms. MoRe also includes a set of state of the art that
with a few lines of code, developers can set up to generate the contextual
recommendations.
• Software architects who want to create a data architecture to support the user,
context and item information for a CARS. To these users MoRe presents an
extensible class structure that supports all this information, and they can extend
or user it straight into the system architecture.
• Researchers of recommendation systems area, who want to compare existing
recommendation algorithms with a new approach, as they can use load their
dataset(s) into MoRe data model and perform recommendations over it using the
contextual algorithms included. Or researchers who just want to put their
algorithms to test in a real scenario, and don’t want to spend time designing a
data model.
A Comprehensive Context-Aware Recommender System Framework 9

Fig. 1 Architectural view of

MoRe framework

MoRe is implemented as a C# Framework following the Object-Oriented

Paradigm (OOP). MoRe can be set up with a few lines of code, and can help in the
modeling and management (storage and retrieval) of user, context, and item
information, as well in the process of generating contextual recommendations based
on the acquired information. The central element of the framework is a
context-aware user model designed specifically for CARS, that along with a data
management module, and with a recommendation algorithms constitute the pro-
posed MoRe framework.
Next, Fig. 1 presents an overall view of MoRe framework architecture, then
each component of the framework is described. Section 3.1 describes the data
model, 3.2 describes de data management, and Sect. 3.3 describes the algorithms
included in MoRe framework.

3.1 The Model

In our previous work [11] a Generic User Model for context-aware Recommender
System (GUMORS) was proposed to solve the problem of not existing an extensive
model that can be used as a reference to structure the user and context aspects inside
CARS. GUMORS provides a large set of user, context, and items working together
to create a CARS specific data model, that can be used into multiple CARS
domains. MoRe uses GUMORS as data structure to manage all the information
needed by prediction functions.
GUMORS organize the CARS information into 4 main top-level categories,
namely: User, Context, Item, and Activity information. Next, the four top-level
categories are briefly described, for a more detailed description please refer to [11].
10 S. Inzunza and R. Juárez-Ramírez

3.1.1 User Aspects

User aspects represent the user as a human being, which CARS can use to infer the
user preferences and behavior to better suit the recommendation results.
User aspects are categorized into Contact, Demographic, Physiological, Interest
and Preferences, Role, Personality, Emotion, and Mental information. Next Table 1
describes each category.

3.1.2 Context Aspects

GUMORS represent the context information in 6 categories, namely: Computing,

Location, Time, Physical Conditions, Resource and Social Relation. Next Table 2
briefly describe each context category.

3.1.3 Activity Information

The activities information that can be useful for CARS are also modeled into
GUMORS. An activity relates information about the User that is performing it, in a
specified Context. A more specialized type of activity is called RatedActivity which
also includes information about the item consumed during the activity, and the
feedback (Rating) the user provided. The activity information is used for example
by Google Maps, to suggest the user navigation route according to his displacement
activity (walking, biking, driving, etc.). Spotify also uses activity information to

Table 1 User aspect considered by GUMORS

Category Category description
Contact Refers to identify the user from the rest of the user in the system, this also
includes information about the user account in the system like the full
name of the user, his email and address
Demographic Describes the demographical information of the user like gender, birthday
and language that the user
Emotion Represents information about the subjective human emotions of the users
Interest and Explicitly describes the interests and preferences of the user for certain
preferences items, and can also be used to store system preferences (e.g. font size)
Mental Used to describe the user’s state of mind, this category includes
information about user’s mood; mental state and cognitive style that can
be used to provide more tailored recommendations
Personality Describes permanent or very slow changing patterns that are associated
with an individual
Physiological Models the aspects of the human body and its functionality
Role This category represents the roles the users play during their activities
A Comprehensive Context-Aware Recommender System Framework 11

Table 2 Contextual information considered by GUMORS

Category Description
Computing This category refers to information about computational elements that the
user is interacting with, or that are blended into the environment
Location Refers to information that relates an item, user, or other context information
with a geographical position. Location include information about the Place
either Physical or Digital, Address and Coordinates for physical
Time Used to capture information about time, such as Date, TimeOfYear,
TimeOfDay, etc.
Physical Describes environmental conditions where the system or user is situated at
condition certain point on time
Resource Model relevant characteristics of the physical or virtual environment
Social Refers to social associations or connections between the user of the system
relations and other persons

Table 3 Classes used by GUMORS to represent the activity information category

Class Description
Activity Main class used to represent the activities. This class has attributes like the Time
Stamp, Name of the activity, and relations to a User and a Context
Rated This class is a specialization of the Activity class. This class includes information
activity of the Item involved in the activity, and the Rating the user gives to such
experience. GUMORS include the following concrete specialization of Rated
Activity: View, Listen, Eat, Travel and Purchase

recommend users songs which rhythm matches his running speed. Next Table 3
shows the classes used by GUMORS to model the activity information.

3.1.4 Items Information

GUMORS also manages the information about the Items the CARS will recom-
mend. For this feature, GUMORS uses an Item super-class that developers can
further specialize through inheritance to fit their specific needs. GUMORS also
contains a series of Item specialization based on the most commonly used items in
CARS literature and considers the commonly used attributes for each item (as
shown in Fig. 2).

3.2 Data Management

The data management module is in charge of persisting the data from the model to a
database and retrieving it back when asked. MoRe uses Entity Framework [33] for
the Object-Relational Mapping (ORM). The use of an ORM for the data persistence
12 S. Inzunza and R. Juárez-Ramírez

Fig. 2 Classes used by GUMORS to model the items of CARS

allows the framework to perform CRUD (Create, Read, Update and Delete) oper-
ations directly over the model classes that is automatically translated into the
database without the need to directly perform queries to the db.
MoRe uses a ModelEntity superclass that contains generic methods with the
logic for all the CRUD operations, using both, synchronous and asynchronous calls.
This superclass is inherited from all the model classes that should be mapped to
database, which inherits the CRUD operations, e.g. Find, FindAll, Save, Delete,
Get, GetAll, etc.

3.2.1 Dataset Generator

The dataset generator is responsible for collecting the data from the database and
organize it so recommendations algorithms can understand it, e.g. Comma
Separated Values (CSV) or Binarized format [34].
MoRe uses Data Annotations to annotate the classes and attributes that will be
part of the resulting dataset. Annotating classes and attributes work as rules that
latter the framework will use to automatically create the dataset, which can be read
by algorithms from memory. As MoRe is designed to work custom internal or
external algorithms, the dataset can also be exported to a file in the specified format,
so any existing algorithm can use the data to generate recommendations.

3.3 Recommendation Algorithms

MoRe uses a large set of algorithms for both, traditional (2D) and contextual
(nD) recommendation. For this first iteration, MoRe uses the set of algorithms
implemented by [28] and a custom wrapper to execute them from C# as originally
the implementation were made under Java development language.
A Comprehensive Context-Aware Recommender System Framework 13

As the framework contained a large set of ready to use algorithms, it can be used
by developers who don’t want to spend too much time implementing existing
recommendation techniques, or by developers that have little experience imple-
menting this type of complex systems [5], the recommendation algorithms can be
used as a black-box, that need to fed whit a dataset (described in previous section),
and will yield a list of recommended item for the specified user.
The framework also allows for custom algorithms to be included, which makes
MoRe ideal for researchers who want to test their new algorithm or recommen-
dation approach against existing state of the art algorithms.
Next, the traditional and contextual recommendation algorithms are described.

3.3.1 Traditional Recommendation Algorithms

Even when MoRe is created to be a context-aware recommendation framework, it

includes traditional recommendation as they are used by filtering recommendation
techniques, described in later section, and even when is not the goal, the framework
can be used for traditional recommendation systems. For this iteration, MoRe
include some of the main traditional recommendation algorithms for every category
(as shown in Table 4), based on the implementation provided by LibRec [35].

3.3.2 Contextual Recommendation Algorithms

The multi-dimensional recommendation algorithms are the core of any recom-

mendation framework, therefore MoRe provides several different algorithms from
state of the art techniques, which gives developers the ability to choose the algo-
rithm that yields better results for their specific domain, as some algorithms may
perform better when recommending movies based on user’s companions, and
others perform better when recommending songs or books.

Table 4 Traditional recommendation techniques and algorithms supported by MoRe

Technique Brief description Algorithms
Ranking This recommendation technique build a SLIM BPR
recommendation recommendation model by analyzing the importance
of each with respect all other items in the system [36]
Average A simple recommendation technique that perform GlobalAvg
recommendation prediction by the average rating given to an element UserAvg
[37] ItemAvg
Collaborative Collaborative-Filtering (CF) techniques construct a UserKNN
filtering model of similarities between users or items based on ItemKNN
recommendation the idea that if two users gave a similar rating for a SVD ++
specific movies, the users or items are very similar
[38]
14 S. Inzunza and R. Juárez-Ramírez

In addition to the list of implemented and ready to work algorithms, MoRe allow
developers to implement their own algorithms o to create hybridizations combining
multiple of the included ones. To implement new algorithms, developers need to
implement the IRecommender or IContextualRecommender class for a new tradi-
tional or contextual algorithms respectively and implement the methods each
interface enforces.
Contextual Filtering Algorithms, try to pre- or post-process the information and
convert it from a multi-dimensional matrix to a 2-dimensional rating matrix that
contains only the user reference (Id), item reference, and rating. Then, this
two-dimensional matrix can be used in traditional algorithms to generate recom-
mendations. There exist many contextual filtering algorithms, one of the most
effective is the context-aware splitting [39]. MoRe includes the three main variants
of this algorithm:
– UserSplitting: From a user point of view, the preference for certain items may
vary depending on the context. User Splitting group items based on the context
and the rating the user gave to them, so when a user gives items good values in a
certain context (e.g. sunny weather) and gives bad rating in another context
situation (e.g. rainy weather), the user can be divided into two users based on
such context (weather in this example).
– ItemSplitting: Separate the item that was rated differently under different context
as being different items. Based on the same rules of UserSplitting to decide
when some instances should be divided, but this process divide items instead of
users.
– UserItemSplitting: Combines both previous separations, the result is that an item
rated by a user in different context, is converted into 2 items, and the user is also
divided into 2 users.
Contextual Modeling Algorithms, unlike splitting approaches, take into account
the influence of context all context aspect on the rating prediction model, which
required the development of new algorithms capable of processing multi-
dimensional matrix and produce a recommendation. The algorithms used for con-
textual modeling supported by MoRe are:
– Tensor Factorization (TF) [40] This algorithm is based on the Matrix
Factorization (MF) dimensional reduction technique which is used for 2D rec-
ommendations. TF consist of extending the two-dimensional MF problem into a
multi-dimensional version, where the rating tensor is converted into a
lower-dimensional vector space. Such that the interactions between users, items,
and contextual factors are represented by a latent factor vector.
– Context-Aware Matrix Factorization (CAMF) [41] is a more scalable (than TF)
contextual modeling approach based on MF. CAMF uses a contextual baseline
prediction function to represent the interactions of contextual information with
the items or users. Baltrunas et al. [42] proposed different variants of CAMF that
model the influence of contextual conditions at different granularities. CAMF-C
assumes that a context factor affects the user and items in the same way.
A Comprehensive Context-Aware Recommender System Framework 15

CAMF-CI models the in influence of a contextual aspect over items. And

CAMF-CC assumes the context affects the ratings for all the items, ignoring the
effect of context over users. MoRe contains the three variants (CAMF-C,
CAMF-CI and CAMF-CC) variants of this algorithm.
Custom Context-based Algorithms, MoRe supports the addition of new algo-
rithms defined by developers and researchers, which allow them to use the
framework to implement and put their theoretical algorithm to test. Any new
contextual algorithm that will be added to MoRe must implement the
IContextAwareRecommender interface, which defines a method to build, save and
load the prediction model, as well as the recommend method what should return a
list of RecommendedItems.

4 Evaluations

Assessing the quality of complex software systems is generally a difficult task,

especially when the software provides new functionality that is not supported by
other software. As previously mentioned, MoRe is the only one of its kind, com-
prehensive Context-Aware Recommender Systems framework. Therefore a direct
comparison with any other proposal can yield inaccurate and misleading results.
To validate MoRe applicability and effectiveness, first, we perform comparative
assessments with other frameworks and libraries’ features. Then some use cases are
implemented using MoRe. For both tests, Microsoft Visual Studio 2017 was used
as development IDE, along with Sql Server 2012 express as database engine, with a
Core i7, 16 GB Ram on Windows 10 computer.

4.1 Comparative Assessment

This section compares MoRe with other framework and libraries discussed in
Sect. 2 of this document.
Next Table 5 presents a comparison between MoRe and other related
context-aware recommender frameworks. The Type section describes whether the
proposal is presented as a theory, a library or a framework; the Data section
describes if the proposal presents a way to structure the information about the user,
context, and items, and if it supports a temporal state profile which maintains a
sub-profile of the user with the current information so it can be quick and easily
accessible by the application. The Algorithm section describes the recommendation
techniques supported by each proposal, and finally, the last section describes the
programming language used by the proposal and if it is available as an open source
(OS), not disclosed (ND) or not available (NA).
16 S. Inzunza and R. Juárez-Ramírez

Table 5 Comparison of recommendation frameworks

MoRe Hybreed MyMediaLite ConRec CoGrec CARSKit
Type
Theoretical •
Library • • •
Framework • •
Data
User model • ◦ ◦ ◦
Context model • ◦ ◦
Item model •
Temporal state • • ◦
Algorithms
Traditional
Ranking • • ◦
Average • • • ◦
Col. filtering • • • ◦
Contextual
Contextual • ◦ ◦
filtering
Contextual • • • • ◦
modeling
Group • •
recommender
Other
Language C# Java C# Java Java
Availability OS OS OS ND NA OS
• Denotes full support; ◦ Partial support; OS: Open Source, ND: Non Disclosed

MoRe is compared against Hybreed [5], MyMediaLite [24], ConRec [1],

CoGrec [26] and CARSKit [28].

4.2 Use Cases

MoRe is framework created to support the creation of context-aware recommender

systems, so is very important the framework work in real-world usages as described
conceptually. To test the applicability of the framework in CARS developments,
and to validate that the resulting framework fulfills the requirements described in
Sect. 2, some use cases where implemented. These use cases represent essential
task of any CARS development (i.e. data storage and retrieval, and performing
recommendations on the stored data).
A Comprehensive Context-Aware Recommender System Framework 17

Fist, the data management feature of the framework is tested selecting a

real-world dataset from literature and loaded into MoRe’s data model. Then, the
ability of the framework to generate recommendations is tested using the loaded
data, and generating predictions using different recommendation techniques.

4.2.1 Data Management Feature

Part of the MoRe’s core is a data structure capable of modeling the users, context,
and items information that will be used by recommendation algorithm to generate
predictions. To test this feature, the LDOS CoMoDa [43] dataset was used loaded
into MoRe’s model. LDOS CoMoDa is a movie dataset that contains 2296
instances, and 32 attributes. The attributes are 9 corresponding to user information,
9 of contextual information and 14 attributes used to describe the items (movies).
Process
The process followed was to set up a new C# project in Visual Studio, and add
references to MoRe framework. As the dataset’s data is separated in 3 files, first the
Items information was loaded into the model by reading the itemsTitles and
itemsMetaData files, then the user, context and ratings information was loaded from
the LDOS-COMODA file. Next Table 6 show the user, context and items aspects
considered in the dataset and how they were mapped to MoRe data model.
Once the data was loaded into the model, and to test the ability of the framework
to serve the data, a dataset was creating using class annotation (as described in
Sect. 3.2) to configure the exportation behavior of the classes. Such recreated
dataset set was used in the next section of use case, where the functionality of the
recommendation algorithms is tested.
Results
The use case of loading data to MoRe’s model showed that the framework was
capable of supporting all LDOS CoMoDa dataset features, most of them (93%)
without the need of any adaptation to the model. Two features (Decision and
Interaction, the first refers to what motivate the user to watch the video, and the
latter refers to the number of interactions of the user with the video) were not
directly supported by the model as shipped with the framework.
To support these two features into the model, some adaptations were required,
the adaptations consisted of further specializing (through inheritance) the Video and
ModelORM classes, adding the required features to the new video (CoMoDaVideo)
class, then registering this CoMoDaVideo class to the specialized ModeORM class.
As the data structure has changed, a database migration was required to synchronize
the data structure with the database structure.
18 S. Inzunza and R. Juárez-Ramírez

Table 6 LDOS CoMoDa dataset attributes mapped to MoRe data model

Category Aspects Supported in
User Id User.Id
Age User.Demographic.Birthday
Gender User.Demographic.Gender
City User.Contact.Address.City
Country User.Contact.Address.Country
EndEmotion User.Emotions.EmotionalState (list)
DominantEmo User.Emotions.EmotionalState (list)
Mood User.Mental.Mood
Physical User.Physiology.PhysiologicalState (list)
Context Id Context.Id
Time Context.Time.TypeOfDay
DayType Context.Time.DayType
Season Context.Time.Season
Location Context.Place.PlaceType
Weather Context.PhysicalCondition.Weather
Social Context.SocialRelation.SocialRelation (list)
*
Decision not directly supported
*
Interaction not directly supported
Item Id Video.Id
Title Video.Title
Director Video.Director
Country Video.Conuntry
Language Video.Language (list)
Year Video.Year
Genre (3) Video.Genre (list)
Actor (3) Video.Actor (list)
Budget Video.ItemAttributes (list)
ImdbUrl Video.ItemAttributes (list)
*
denotes attributes not directly supported in the model, some adaptation was required

Discussion
The selected dataset can be seen as having a small number of features (32), but most
of the existing dataset for CARS have fewer features (e.g. InCarMusic [41] as 18,
DePaul Movie [28] has 6, and Trip Advisor [39] has 9). Even though, being able to
use a bigger dataset is an interesting challenge that would allow us to test more
attributes of the model, having used LDOS CoMoDa dataset allow us to test the
ability of the model to store data, to maintain the relationship between the data
categories (user, context, and item), and the ability to serve back the stored data.
The age attribute contained in the dataset was considered supported even when it
was not stored directly as an integer, rather the model stores the birthday that which
can be easily converted into age with a little processing. Maybe such processing of
converting birthday into age should be directly included in the model in a future
release of MoRe. The emotions (endEmotion and dominantEmo) and mood are
considered as part of the user information, but as they are intended to reflect the
user’s state of mind in a specific situation, they are linked to the context, as the
A Comprehensive Context-Aware Recommender System Framework 19

MoRe models support to assign a specific context to emotional situation, meaning

that the stored emotion was triggered/reading from the specified (if any) context.

4.2.2 Contextual Recommendation Feature

To test the ability of MoRe framework to generate context-based recommendations,

which is the core functionality of the framework, the following use case was
implemented.
Process
The process for this experiment consisted of using the framework’s dataset gen-
erator to create the data in the database into a data matrix that can be fed to the
algorithms, to generate recommendations, simulating the process that will be fol-
lowed in a real-world CARS implementation.
The data used for this experiment was LDOS CoMoDa dataset that was loaded
onto the framework model (as described in previous section). Having the dataset
created, it was used to generate predictions using the both, contextual filtering and
contextual modeling techniques.
Typically, algorithms are evaluated on their rating prediction ability, using
metrics like Mean Absolute Error (MAE) and Root Mean Square Error (RMSE),
such metrics were used to evaluate MoRe algorithms, except for SLIM and BPR
algorithms, are they only support Top-N item recommendation, and not rating
prediction, these algorithms were evaluated using ranking metrics Area Under the
Curve (AUC) and Recall for 5 elements (Rec5).
In this experiment, all the techniques and algorithms of MoRe were tested,
including the 24 combinations of the three splitting approaches (UserSplitting,
ItemSplitting and UISplitting) with the 8 traditional recommendation algorithms
(SLIM, BPR, GlobalAverage, UserAverage, ItemAverage, SVD++, UserKNN and
ItemKNN).
Results
The obtained results loading the LDOS CoMoDa dataset into MoRe, generating a
dataset with such information and applying the recommendation algorithms are
shown next. Figure 3 shows the RMSE and MAE values for the contextual mod-
eling algorithms; then Table 7 shows the result of the combination of filtering
techniques with the baseline recommender algorithms.
Discussion
This experiment was used as a proof of concept and allows us to test the func-
tionality of one of the main features MoRe provides: generating recommendations
over stored data. All the algorithms contained in MoRe were testes with the same
dataset, yielding somewhat similar results. The better performing algorithm was
ItemAverage when used with ItemSplitting approach. In general, the contextual
20 S. Inzunza and R. Juárez-Ramírez

Fig. 3 Evaluation of contextual modeling algorithms on LDOS CoMoDa data

Table 7 Evaluation of contextual filtering algorithms on LDOS CoMoDa data

Technique Algorithm UserSplitting ItemSplitting UISplitting
AUC Rec5 AUC Rec5 AUC Rec5
Ranking SLIM 0.5522 0.0710 0.5507 0.6800 0.5526 0.0730
BPR 0.5029 0.0030 0.5019 0.0030 0.5036 0.0039
MAE RMSE MAE RMSE MAE RMSE
Collaborative filtering GlobalAvg 0.7700 0.9789 0.8509 1.0590 0.8509 1.0580
UserAvg 0.7700 0.9789 0.7615 0.9757 0.7697 0.9788
ItemAvg 0.4728 0.7461 0.4870 0.0770 0.4969 0.7701
Average SVD ++ 0.8573 1.0662 0.8561 1.0686 0.8552 1.0643
UserKNN 0.8125 1.0447 0.8173 1.0501 0.8210 1.0502
ItemKNN 0.7041 0.9272 0.7089 0.9331 0.7170 0.9388
Bold values are the best performing results

splitting approach yield slightly better results, especially when paired with
ItemAverage baseline recommender.
The obtained results not necessary means that this will be the better performing
algorithms in all cases, as these results are specific to recommend movies, and
taking in consideration the specific characteristics of the LDOS CoMoDa dataset
(user and context aspects considered, the data sparsity, and the number of instan-
ces). When recommending other elements, and using other contextual information,
the algorithms will perform differently, fortunately, MoRe contains a large set of
algorithms so developers can use the better performing one in their specific CARS.

5 Conclusions and Future Work

In this document, we have introduced MoRe, a comprehensive software framework

to build context-aware recommender systems. To our knowledge, MoRe is the first
CARS framework that provides developers a context-aware user model structure
Exploring the Variety of Random
Documents with Different Content
liued here vpon earth. And therefore it is said by Plato and other,
that Orpheus after his death had his soule thrust into the bodie of a
swanne, that of Agamemnon conueied into an egle, of Aiax into a
lion, of Atlas into a certeine wrestler, of Thersites into an ape, of
Deiphobus into Pythagoras, and Empedocles dieng a child, after
sundrie changes into a man, whereof he himselfe saith;

Ipse ego námq; fui puer olim, deinde puella,

Arbustum & volucris, mutus quóq; in æquore
piscis.

For said they (of whom Pythagoras also had,

Plinius, lib. 16. cap. ultimo.
and taught this errour) if the soule apperteined at the first to a king,
and he in this estate did not leade his life worthie his calling, it
should after his decease be Metempsuchôsis. shut vp in the bodie of a
slaue, begger, cocke, owle, dog, ape, horsse, asse, worme, or
monster, there to remaine as in a place of purgation and
punishment, for a certeine period of time. Beside this, it should
peraduenture susteine often translation from one bodie vnto another,
according to the quantitie and qualitie of his dooings here on earth,
till it should finallie be purified, and restored againe to an other
humane bodie, wherein if it behaued it selfe more orderlie than at
the first: after the next death, it should be preferred, either to the
bodie of a king againe, or other great estate. And thus they made a
perpetuall circulation or reuolution of our soules, much like vnto the
continuall motion of the heauens, which neuer stand still, nor long
yeeld one representation and figure. For this cause also, as Diodorus
saith, they vsed to cast certeine letters into the fire, wherein the
dead were burned, to be deliuered vnto their deceased fréends,
whereby they might vnderstand of the estate of such as trauelled
here on earth in their purgations (as the Moscouits doo write vnto S.
Nicholas to be a speach-man for him that is buried, in whose hand
they bind a letter, and send him with a new paire of shooes on his
feet into the graue) and to the end that after their next death they
should deale with them accordinglie, and as their merits required.
They brought in also the worshipping of manie gods, and their
seuerall euen to this daie Oke honored whereon mistle did grow, and so doo our
sacrifices:
sorcerers thinking some spirits to deale about ye same, for hidden treasure.
they honoured likewise the oke, whereon the mistle groweth, and
dailie deuised infinit other toies (for errour is neuer assured of hir
owne dooings) whereof neither Samothes, nor Sarron, Magus, nor
Druiyus did leaue them anie prescription.

These things are partlie touched by Cicero, Strabo, Plinie, Sotion,

Laertius, Theophrast, Aristotle, and partlie also by Cæsar, Mela, Val.
Max. lib. 2. and other authors of later time, who for the most part
doo confesse, that the cheefe schoole of the Druiydes was holden
here in Britaine, where that religion (saith Plinie) was so hotlie
professed and followed, "Vt dedisse Persis videri possit," lib. 30. cap.
1. and whither the Druiydes also themselues, that dwelt among the
Galles, would often resort to come by the more skill, and sure
vnderstanding of the mysteries of that doctrine. And as the Galles
receiued their religion Logike and Rhetorike out of Gallia.] from the Britons,
so we likewise had from them some vse of Logike & Rhetorike, such
as it was which our lawiers practised in their plees and common
causes. For although the Greeks were not vnknowne vnto vs, nor we
to them, euen from the verie comming of Brute, yet by reason of
distance betwéene our countries, we had no great familiaritie and
common accesse one vnto another, till the time of Gurguntius, after
whose entrance manie of that nation trauelled hither in more
securitie, as diuers of our countriemen did vnto them without all
danger, to be offered vp in sacrifice to their gods. That we had the
maner of our plees also out of France, Iuuenal is a witnesse, who
saith;

Gallia causidicos docuit facunda Britannos.

Howbeit as they taught vs Logike and Rhetorike, so we had also

some Sophistrie from them; but in the worst sense: for from France
is all kind of forgerie, corruption of maners, and craftie behauiour
not so soone as often transported into England. And albeit the
Druiydes were thus honored and of so great authoritie in Britaine,
yet were there great numbers of them also in the Iles of Wight,
Anglesey, and the Orchades, in which they held open schooles of
their profession, aloofe as it were from the resort of people, wherein
they studied and learned their songs by heart. Howbeit the cheefe
college of all I say, remained still in Albion, whither the Druiydes of
other nations also (beside the Galles) would of custome repaire,
when soeuer anie controuersie among them in matters of religion
did happen to be mooued. At such times also the rest were called
out of the former Ilands, whereby it appeareth that in such cases
they had their synods and publike meetings, and therevnto it grew
finallie into custome, and after that a prouerbe, euen in variances
falling out among the princes, great men, and common sorts of
people liuing in these weast parts of Europe, to yeeld to be tried by
Britaine and hir thrée Ilands, bicause they honoured hir préests (the
Druiydes) as the Athenians did their Areopagites.

Furthermore, in Britaine, and

Estimation of the Druiydes or Druiysh preests.
among the Galles, and to say the truth, generallie in all places where
the Druiysh religion was frequented, such was the estimation of the
préests of this profession, that there was little or nothing doone
without their skilfull aduise, no not in ciuil causes, perteining to the
regiment of the common-wealth and countrie. They had the charge
also of all sacrifices, publike and priuate, they interpreted oracles,
preached of religion, and were neuer without great numbers of
yoong men that heard them with diligence, as they taught from time
to time.

Touching their
Immunitie of the cleargie greater vnder idolatrie than vnder the gospell.
persons also they were exempt from all temporall seruices,
impositions, tributes, and exercises of the wars: which immunitie
caused the greater companies of scholers to flocke vnto them from
all places, & to learne their trades. Of these likewise, some remained
with them seuen, eight, ten, or twelue years, still learning the
secrets of those unwritten mysteries by heart, which were to be had
amongst them, and commonlie pronounced in verse. And this
policie, as I take it, they vsed onelie to preserue their religion from
contempt, whereinto it might easilie haue fallen, if any books thereof
had happened into the hands of the common sort. It helped also not
a little in the exercise of their memories, wherevnto bookes are vtter
enimies, insomuch as he that was skilfull in the Druiysh religion,
would not let readilie to rehearse manie hundreds of verses togither,
and not to faile in one tittle, in the whole processe of this his
laborious repetition. But as they dealt in this order for matters of
their religion, so in ciuill affairs, historicall treatises, and setting
downe of lawes, they vsed like order and letters almost with the
Grecians. Whereby it is easie to be séene, that they reteined this
kind of writing from Druiyus (the originall founder of their religion)
and that this Iland hath not béene void of letters and learned men,
euen sith it was first inhabited. I would ad some thing in particular
also of their apparell, but sith the dealing withall is nothing profitable
to the reader, I passe it ouer, signifieng neuerthelesse, that it was
distinguished by sundrie deuises from that of the common sort, and
of such estimation among the people, that whosoeuer ware the
Druiysh weed, might walke where he would without any harme or
annoiance. This honour was giuen also vnto the préests in Rome,
insomuch that when Volusius was exiled by the Triumuirate, and saw
himselfe in such danger, as that he could not escape the hardest, he
gat the wéed of a preest upon his backe, and begged his almes
therein, euen in the high waies as he trauelled, and so escaped the
danger and the furie of his aduersaries: but to proceed with other
things.

Bardus. After the death of Druiyus, Bardus his sonne, and fift king of
the Celts, succéeded not onelie ouer the said kingdome, but also in
his fathers vertues, whereby it is verie likelie, that the winding and
wrapping vp of the said religion, after the afore remembred sort into
verse, was first deuised by him, for he was an excellent poet, and no
lesse indued with a singular skill in the practise and speculation of
musicke, of which two many suppose him to be the verie author and
Gen. 4. 21. beginner, although vniustlie, sith both poetrie and song
were in vse before the flood, as was also the harpe and pipe, which
Iubal inuented, and could neuer be performed without great skill in
musicke. But to procéed, as the cheefe estimation of the Druiydes
remained in the end among the Britons onelie, for their knowledge
in religion, so did the fame of the Bardes (which were so called of
this Bardus for their excellent skill in musicke, poetrie, and the
heroicall kind of song, which at the first conteined onelie the high
mysteries and secret points of their religion. There was little
difference also betwéene them and The Bards degenerate. the Druiydes,
till they so farre degenerated from their first institution, that they
became to be minstrels at feasts, droonken meetings, and
abhominable sacrifices of the idols: where they sang most
commonlie no diuinitie as before, but the puissant acts of valiant
princes, and fabulous narrations of the adulteries of the gods. Certes
in my time this fond vsage, and thereto the verie name of the
Bardes, are not yet extinguished among the Britons of Wales, where
they call their poets and musicians Barthes, as they doo also in
Ireland: which Sulpitius also writing to Lucane remembreth, where
he saith that the word Bardus is meere Celtike, and signifieth a
singer. Howbeit the Romans iudging all nations beside themselues to
be but rude and barbarous, and thereto misliking vtterlie the rough
musicke of the Bardes, entred so farre into the contemptuous
mockage of their melodie, that they ascribed the word Bardus vnto
their fooles and idiots, whereas contrariwise the Scythians and such
as dwell within the northweast part of Europe, did vse the same
word in verie honourable maner, calling their best poets and heroicall
singers, Singebardos; their couragious singers and capiteins that
delited in musicke, Albardos, Dagobardos, Rodtbardos, & one lame
musician Lambard aboue all other, of whose skilfull ditties Germanie
is not vnfurnished, as I heare vnto this daie. In Quizqueia or new
Spaine, an Iland of the Indies, they call such men Boitios, their rimes
Arcitos, and in steed of harps they sing vnto timbrels made of shels
such sonnets and ditties as either perteine vnto religion, prophane
loue, commendation of ancestrie, and inflammation of the mind vnto
Mars, whereby there appeareth to be small difference betwéene
their Boitios and our Bardes. Finallie of our sort, Lucane in his first
booke writeth thus, among other like saiengs well toward the latter
end;
Lucani. li. 1.

Vos quóq; qui fortes animas, bellóq; peremptas

Laudibus in longum vates dimittitis æuum,
Plurima securi fudistis carmina Bardi.
Et vos barbaricos ritus, morémque sinistrum
Sacrorum Druiydæ, positis recepistis ab armis.
Solis nosse Deos, & cœli numina vobis,
Aut solis nescire datum: nemora alta remotis
Incolitis lucis. Vobis authoribus, vmbræ
Non tacitas Erebi sedes, Ditisque profundi
Pallida regna petunt, regit idem spiritus artus
Orbe alio. Longæ canitis si cognita, vitæ
Mors media est, certe populi, quos despicit
arctos,
Fœlices errore suo, quos ille timorum
Maximus haud vrget leti metus: inde ruendi
In ferrum mens prona viris, animæque capaces
Mortis: & ignauum est redituræ parcere vitæ.

Thus we sée as in a glasse the state of religion, for a time, after the
first inhabitation of this Iland: but how long it continued in such
soundnesse, as the originall authors left it, in good sooth I cannot
say, yet this is most certeine, that after a time, when Albion arriued
here, the religion earst imbraced fell into great decaie. For whereas
Iaphet & Samothes with their children taught nothing else than such
doctrine as they had learned of Noah: Cham the great grandfather
of this our Albion, and his disciples vtterlie renouncing to follow their
steps, gaue their minds wholie to seduce and lead their hearers
headlong vnto all error. Whereby his posteritie not onelie corrupted
this our Iland, with most filthie trades and practises; but also all
mankind, generallie where they became, with vicious life, and most
vngodlie conuersation. What doctrine Cham and his disciples taught. For from
Cham and his successours procéeded at the first all sorcerie,
witchcraft, and the execution of vnlawful lust, without respect of sex,
age, consanguinitie, or kind: as branches from an odious and
abhominable root, or streames deriued from a most filthie and
horrible stinking puddle. Howbeit, & notwithstanding all these his
manifold lewdnesses, such was the follie of his Ægyptians (where he
first reigned and taught) that whilest he liued they alone had him in
great estimation (whereas other nations contemned and abhorred
him for his wickednesse, calling Chemesenua. him Chemesenua, that is,
the impudent, infamous and wicked Cham) and not
Chem Min.
Cham made a god. onelie builded a citie vnto him which they called
Chem Min, but also after his death reputed him for a god, calling the
highest of the seuen planets after his name, as they did the next
beneath it after Osyris his sonne, whom they likewise honored vnder
the name of Iupiter.

Certes it was a custome

Translation of mortall men into heauen how it began.
begonne in Ægypt of old time, and generallie in vse almost in euerie
place in processe of time (when any of their famous worthie princes
died) to ascribe some forme or other of the stars vnto his person, to
the end his name might neuer weare out of memorie. And this they
called their translation in heauen, so that he which had any starres
or forme of starres dedicated vnto him, was properlie said to haue a
seat among the gods. A toie much like to the catalog of Romish
saints, (although the one was written in the celestiall or immateriall
orbes, the other in sheeps skins, and verie brickle paper) but yet so
estéemed, that euerie prince would oft hazard and attempt the
vttermost aduentures, thereby to win such fame in his life, that after
his death he might by merit haue such place in heauen, among the
shining starres. Howbeit, euerie of those that were called gods,
could not obteine that benefit, for then should there not haue béene
stars enow in heauen to haue serued all their turnes, wherfore
another place was in time imagined, where they reigned that were
of a second calling, as the Cyril, aduersus Iul. lib. 6. sect. 8. Semones who
were gods by grace and fauour of the people. "Semones dici
voluerunt (saith Fulgentius In vocibus antiquis) quos cœlo nec
dignos ascriberent, ob meriti paupertatem; sicut Priapus Hyppo.
Vortumnus, &c. nec terrenos eos deputare vellent per gratiæ
venerationem," as also a third place that is to say an earth, where
those gods dwelled which were noble men, officers, good
gouernours and lawgiuers to the people, and yet not thought
worthie to be of the second or first companie, which was a iollie
diuision.

Thus we sée in generall maner, how idolatrie, honoring of the

starres, and brood of inferiour gods were hatched at the first, which
follies in processe of time came also into Britaine, as did the names
of Saturne & Iupiter, &c: as shall appeare hereafter. And here sith I
haue alreadie somewhat digressed from my matter, I will go yet a
little farder, and shew foorth the originall vse of the word Saturne,
Iupiter, Hercules, &c: whereby your Honor shall sée a little more into
the errours of the Gentils, and not onelie that, but one point also
Which were properlie called Saturni, Ioues, Iunones, and Hercules. of the root of all
the confusion that is to be found among the ancient histories. Certes
it was vsed for a few yéeres after the partition of the earth (which
was made by Noah, in the 133. yeere after the floud) that the
beginners of such kingdoms as were then erected should be called
Saturni, whereby it came to passe that Nimbrote was the Saturne of
Babylon: Cham of Ægypt: and so foorth other of sundrie other
countries. Their eldest sonnes also that succeeded them, were called
Ioues; and their nephewes or sonnes sonnes, which reigned in the
third place Hercules, by which meanes it followed that euerie
kingdome had a Saturne, Iupiter and Hercules of hir owne, and not
from anie other.

In like sort they had such another order among their daughters,
whom they married as yet commonlie vnto their brethren (God
himselfe permitting the same vnto them for a time) as before the
floud, to the end the earth might be thoroughlie replenished, and
the sooner furnished with inhabitants in euerie part therof. The sister
therefore and wife of Isis, Io and Iuno all one. euerie Saturne was called
Rhea, but of Iupiter, Iuno, Isis, or Io. Beyond these also there was
no latter Harold that would indeuour to deriue the petigree of any
prince, or potentate, but supposed his dutie to be sufficientlie
performed, when he had brought it orderlie vnto some Saturne or
other, wherat he might cease, and shut vp all his trauell. They had
likewise this opinion grounded amongst them, that heauen & earth
were onlie parents vnto Saturne and Rhea, not knowing out of
doubt, Cœlum or Cœlus.
Ogyges.
Sol.
Pater what they themselues did meane, sith these
deorum.
denominations, Heauen, Ogyges, the Sunne, Pater Deorum, and
such like, were onelie ascribed vnto Noah: as *Terra, (the Earth)
Vesta, Aretia, the Moone, Mater * Tydea.
Vesta.
Terra.
Luna.
Aretia.
Deorum mater. deorum, and other the like were vnto Tydea his wife. So
that hereby we sée, how Saturne is reputed in euerie nation for their
oldest god, or first prince, Iupiter for the next, and Hercules for the
third. And therefore sith these names were dispersed in the
beginning ouer all, it is no maruell that there is such confusion in
ancient histories, and the dooings of one of them so mixed with
those of another, that it is now impossible to distinguish them in
sunder. This haue I spoken, to the end that all men may see what
gods the Pagans honored, & thereby what religion the posteritie of
Cham did bring ouer into Britaine. For vntill their comming, it is not
likelie that anie grosse idolatrie or superstition did enter in among
vs, as deifieng of mortall men, honoring of the starres, and erection
of huge images, beside sorcerie, witchcraft, and such like, whereof
the Chemminites are worthilie called the autors. Neither were these
errors anie thing amended, by the comming Frō whence Brute did learne his
religion. in of Brute, who no doubt added such deuises vnto the same,
as he and his companie had learned before in Græcia, from whence
also he brought Helenus the sonne of Priamus, (a man of excéeding
age) & made him his préest and bishop thorough out the new
conquest, that he had atchieued in Britaine.
After Brute, idolatrie and superstition still increased more and more
among vs, insomuch that beside the Druiysh and Bardike
ceremonies, and those also that came in with Albion and Brute
himselfe: our countriemen either brought hither from abroad, or
dailie inuented at home new religion and rites, whereby it came to
passe that in the stead of the onelie and immortall God (of whom
Samothes and his posteritie did preach Dis or Samothes made a god. in
times past) now they honored the said Samothes himselfe vnder the
name of Dis and Saturne: also Iupiter, Mars, Minerua, Mercurie,
Apollo, Diana; and finallie Hercules, vnto whome they dedicated the
gates and porches of their temples, entrances into their regions,
cities, townes and houses, with their limits and bounds (as the
papists did the gates of their cities and ports vnto Botulph & Giles)
bicause fortitude and wisedome are the cheefe vpholders and
bearers vp of common-wealths and kingdoms, both which they
ascribed to Hercules (forgetting God) and diuers other idols whose
names I now remember not. In lieu moreouer of sheepe and oxen,
Mela. Diodorus, Strab. 4. Plin. Cæsar. 5. they offred mankind also vnto some
of them, killing their offendors, prisoners, and oft such strangers as
came from farre vnto them, by shutting vp great numbers of them
togither in huge images made of wicker, réed, haie, or other light
matter: and then setting all on fire togither, they not onelie
consumed the miserable creatures to ashes (sometimes adding other
beasts vnto them) but also reputed it to be the most acceptable
sacrifice that could be made vnto their idols. From whence they had
this horrible custome, trulie I cannot tell, but that it was common to
most nations, not onlie to consume their strangers, captiues, &c; but
also their owne children with fire, in such maner of sacrifice: beside
the text of the Bible, the prophane histories doo generallie leaue it
euident, as a thing either of custome or of particular necessitie, of
which later Virgil saith;

Sanguine placastis ventos & virgine cæsa, &c.

As Silius dooth of the first, where he telleth of the vsuall maner of

the Carthaginenses, saieng after this maner;
Vrna reducebat miserandos annua casus, &c.

But to procéed with our owne gods and idols, more pertinent to my
purpose than the rehersall of forreine demeanours: I find that huge
temples in like sort were builded vnto them, so that in the time of
Lucius, when the light of saluation began stronglie to shine in Ptol.
Lucensis. Britaine,thorough the preaching of the gospell, the christians
discouered 25. Flamines or idol-churches beside three Archflamines,
whose préests were then as our Archbishops are now, in that they
had superior charge of all the rest, the other being reputed as
inferiours, and subiect to their iurisdiction in cases of religion, and
superstitious ceremonies.

Of the quantities of their idols I speake not,

Monstrous proportions of idols.
sith it is inough to saie, that they were monstrous, and that each
nation contended which should honour the greater blocks, and yet
all pretending to haue the iust heigth of the god or goddesse whom
they did represent. Apollo Capitolinus that stood at Rome, was thirtie
cubits high at the least; Tarentinus Iupiter of 40.; the idoll of the
sonne in the Rhodes, of 70 (whose toe few men could fadam;)
Tuscanus Apollo that stood in the librarie of the temple of Augustus,
of 50. foot; another made vnder Nero of 110. foot; but one in France
passed all, which Zenoduris made vnto Mercurie at Aruernum in ten
years space, of 400. foot. Wherby it appeareth, that as they were
void of moderation in number of gods, so without measure were
they also in their proportions, and happie was he which might haue
the greatest idoll, and lay most cost thereon.

Hitherto yee haue heard of the time, wherein idolatrie reigned and
blinded the harts of such as dwelled in this Iland. Now let vs sée the
successe of the gospell, after the death and passion of Iesus Christ
our sauiour. And euen here would I begin with an allegation of
Theodoret. Theodoret, wherevpon some repose great assurance
(conceiuing yet more Sophronius. hope therein by the words of
Sophronius) that Paule the Apostle should preach the word of
saluation here, after his deliuerie out of captiuitie, which fell as I doo
read in the 57. of Christ. But sith I cannot verifie the same by the
words of Theodoret, to be spoken more of Paule than Peter, or the
rest, I will passe ouer this coniecture (so far as it is grounded vpon
Theodoret) and deale with other authorities, whereof we haue more
certeintie. First of all therfore let vs see what Fortunatus hath written
of Pauls comming into Britaine, and afterward what is to be found of
other by-writers in other points of more assurance. Certes for the
presence of Paule I read thus much:

Quid sacer ille simul Paulus tuba gentibus

ampla,
Per mare per terras Christi præconia fundens,
Europam & Asiam, Lybiam, sale dogmata
complens,
Arctos, meridies, hic plenus vesper & ortus,
Transit & Oceanum, vel qua facit insula
portum,
Quásq; Britannus habet terras atque vltima
Thule, &c.

Iosephus.That one Iosephus preached here in England, in the time of

the Apostles, his sepulchre yet in Aualon, now called Glessenburg or
Glastenburie, an epitaph affixed therevnto is proofe sufficient.
Howbeit, sith these things are not of competent force to persuade all
men, I will ad in few, what I haue read elsewhere of his arriuall
here. First of all therefore you shall note that he came ouer into
Britaine, about the 64. after Christ, when the persecution began
vnder Nero, at which time Philip and diuers of the godlie being in
France (whether he came with other christians, after they had
sowed the word of God in Scythia, by the space of 9. yeares)
seuered themselues in sunder, to make the better shift for their
owne safegard, and yet not otherwise than by their flight, the
gospell might haue due furtherance. Hereby then it came to Philip.
Freculphus. To. 2., lib. 2. cap. 4.
Nennius. Nicephorus lib. 2. cap. 40.
Isidorus lib. de vita & obit. dict. patrum.
W. Malmes. de antiq. Glasconici monast. passe, that the said Philip vpon good
deliberation did send Iosephus ouer, and with him Simon Zelotes to
preach vnto the Britons, and minister the sacraments there
according to the rites of the churches of Asia and Greece, from
whence they came not long before vnto the countrie of the Galles.
Which was saith Malmesburie 103. before Faganus and Dinaw did
set foorth the gospell amongst them. Of the cōming of Zelotes you
may read more in the second booke of Niceph. Cal. where he writeth
thereof in this maner: "Operæpretium etiam fuerit Simonem Cana
Galileæ ortum, qui propter flagrantem in magistrum suum ardorem,
summámq; euangelicæ rei per omnia curam Zelotes cognominatus
est hîc referre, accepit enim is cœlitùs adueniente spiritu sancto,
Aegyptium Cyrenem & Africam, deinde Mauritaniam & Lybiam
omnem euangelium deprædicans percurrit, eandemque doctrinam
etiam ad occidentalem Oceanum insulásque Britannicas perfert." And
this is the effect in a little roome, of that which I haue read at large
in sundrie writers, beside these two here alledged, although it may
well be gathered that diuers Britains were conuerted to the faith,
before this sixtie foure of Christ. Howbeit, whereas some write that
they liued, and dwelled in Britaine, it cannot as yet take any absolute
hold in my iudgement, but rather that they were baptised and
remained, either in Rome, or else-where. And of this sort I suppose
Claudia Rufina a British ladie. Claudia Rufina the wife of Pudens to be one,
who was a British ladie indeed, and not onelie excellentlie séene in
the Gréeke and Latine toongs, but also with hir husband highlie
commended by S. Paule, as one 1. Tim. 4. hauing had conuersation
and conference with them at Rome, from whence he did write his
second epistle vnto Timothie, as I read. Of this ladie moreouer
Martial speaketh, in reioising that his poesies were read also in
Britaine, and onelie by hir meanes, who vsed to cull out the finest &
honestest of his epigrams and send them to hir fréends for tokens,
saieng after this maner, as himselfe dooth set it downe:

Dicitur & nostros cantare Britannia versus.

Furthermore making mention of hir and hir issue, he addeth these
words:

Li. 11. Epig. 54.

Claudia cœruleis cùm sit Rufina Britannis

Edita, cur Latiæ pectora plebis habet?
Quale decus formæ? Romanam credere matres
Italides possunt, Atthides esse suam.
Dij bene, quod sancto peperit fæcunda marito,
Quot sperat, generos, quótque puella nurus.
Sic placeat superis, vt coniuge gaudeat vno,
Et semper natis gaudeat illa tribus.

The names of hir thrée children were Prudentiana, Praxedes, both

virgins, and Nouatus, who after the death of Pudens their father
(which befell him in Cappadocia) dwelled with their mother in
Vmbria, where they ceased not from time to time to minister vnto
the saints. But to leaue this impertinent discourse, and proceed with
my purpose.

I find in the Chronicles of Burton (vnder the yeare of Grace 141. and
time of Hadrian the emperour) that nine scholers or clerkes of
Grantha or Granta (now Cambridge) were baptised in Britaine, and
became preachers of the gospell there, but whether Taurinus bishop
or elder ouer the congregation at Yorke (who as Vincentius saith,
was executed Lib. 10. cap. 17.
Taurinus. about this time for his faith) were one of them or not, as yet
I do not certeinlie find; but rather the contrarie, which is that he was
no Britaine at all, but Episcopus Ebroicensis, for which such as
perceiue not the easie corruption of the word, may soone write
Eboracensis as certeinlie mine author out of whom I alledge this
authoritie hath done before me. For Vincentius saith flat otherwise,
and therefore the Chronologie if it speake of anie Taurinus bishop of
Yorke is to be reformed in that behalfe. Diuers other also imbraced
the religion of Christ verie zealouslie before these men. Howbeit, all
this notwithstanding, the glad tidings of the gospell had neuer free
and open passage here, vntill the time of Lucius, in which the verie
enimies of the word became the apparent meanes (contrarie to their
owne minds) to haue it set foorth amongst vs. For when Antoninus
the emperour had giuen out a decrée, that the Druiysh religion
should euerie where be abolished, Lucius the king (whose surname
is now perished) tooke aduise of his councell what was best to be
doone, & wrote in this behalfe. And this did Lucius, bicause he knew
it *impossible for man to liue long * This is contrarie to the common talke of our
Atheists who say, Let vs liue here in wealth, credit and authoritie vpon earth, and let God
without any
take heauen and his religion to himselfe to doo withall what he listeth.
religion at all: finallie finding his Nobilitie & subiects vtter enemies to
the Romane deuotiō (for that they made so many gods as they
listed, & some to haue the regiment euen of their dirt & dung) and
thervnto being pricked forwards by such christians as were
conuersant about him, to choose the seruice of the true God that
liueth for euer, rather than the slauish seruitude of any pagan idoll:
he fullie resolued with himselfe in the end, to receiue and imbrace
the gospell of Christ. He sent also two of his best learned and
greatest Lucius openeth his ears to good counsell, as one desirous to serue God & not
prefer the world. philosophers to Rome, vnto Eleutherus then bishop
there in the 177. of Christ, not to promise any subiection to his sea,
which then was not required, but to say with such as were pricked in
mind, Acts. 2. verse. 37. "Quid faciemus viri fratres?" I meane that
they were sent to be perfectlie instructed, and with farther
commission, to make earnest request vnto him and the congregation
there, that a competent number of preachers might be sent ouer
from thence, by whose diligent aduise and trauell, the foundation of
the gospell might surelie be laid ouer all the portion of the Ile, which
conteined his kingdome, according to his mind.

When
The purpose of Lucius opened vnto the congregation at Rome by Eleutherus.
Eleutherus vnderstood these things, he reioiced not a little for the
great goodnesse, which the Lord had shewed vpon this our Ile and
countrie. Afterwards calling the brethren togither, they agréed to
ordeine, euen those two for bishops, whom Lucius as you haue
heard, had directed ouer vnto them. Finallie after they had
thoroughlie catechized them, making generall praier vnto God and
earnest supplication for the good successe of these men, they sent
them home againe with no small charge, that they should be diligent
in their function, and carefull ouer the flocke committed to their
custodie.

The first of these was called Eluanus Aualonius, a man borne in the
Ile of Aualon, and brought up there vnder those godlie pastours and
their disciples, whom Philip sent ouer at the first for the conuersion
of the Britons. The other hight Medguinus, and was thereto
surnamed Belga, bicause he was of the towne of Welles, which then
was called Belga. This man was trained vp also in one schoole with
Eluanus, both of them being ornaments to their horie ages, and men
of such grauitie and godlinesse, that Eleutherus supposed none
more worthie to support this charge, than they: after whose
comming home also, it was not long yer Lucius and all A zealous prince
maketh feruent subiects. his houshold with diuers of the Nobilitie were
baptised, beside infinit numbers of the common people, which dailie
resorted vnto them, and voluntarilie renounced all their idolatrie and
paganisme.

In the meane time, Eleutherus vnderstanding the successe of these

learned doctours, and supposing with himselfe, that they two onlie
could not suffice to support so great a charge as should concerne
the conuersion of the whole Iland; he directed ouer vnto them in the
yeare Faganus.
Dinauus.
Aaron. insuing Faganus, Dinaw (or Dinauus) Aaron, and diuerse other
godlie preachers, as fellow-labourers to trauell with them in the
vineyard of Radulphus de la noir aliàs Niger. the Lord. These men therefore
after their comming hither, consulted with the other, and foorthwith
wholie consented to make a diuision of this
3. Cheefe Bishops in Britaine. Iland amongst themselues, appointing what
parcell each preacher should take, that with the more profit and
ease of the people, and somewhat lesse trauell also for themselues,
the doctrine of the Gospell might be preached and receiued. In this
distribution, they ordeined that there Theonus.
Theodosius.
London.
Yorke.
Caerlheon. should be one congregation at London, where they placed
Theonus as chéefe elder and bishop, for that present time, worthilie
called Theonus. 1. for there was another of that name who fled into
Wales with Thadiocus of Yorke, at the first comming of the Saxons;
and also Guthelmus, who went (as I read) into Armorica, there to
craue aid against the Scots and Vandals that plagued this Ile, from
the Twede vnto the Humber. After this Theonus also Eluanus
succéeded, who conuerted manie of the Druiydes, and builded the
first librarie neere vnto the bishops palace. The said Lucius also
placed another at Yorke, whither they appointed Theodosius: and
the third at Caerlheon vpon the riuer Vske, builded sometimes by
Belinus, and called Glamorgantia, but now Chester (in which three
cities there had before time beene thrée Archflamines erected vnto
Apollo, Mars, and Minerua, but now raced to the ground, and three
other churches builded in their steeds by Lucius) to the end that the
countries round about might haue indifferent accesse vnto those
places, and therewithall vnderstand for certeintie, whither to resort
for resolution, if after their conuersion they should happen to doubt
of any thing. In like sort also the rest of the idoll-temples standing in
other places were either ouerthrowne, or conuerted into churches
for christian congregations to assemble in, as our writers doo
remember. In the report whereof giue me leaue gentle reader, of
London my natiue citie to speake a little: for although it may and
dooth seeme impertinent to my purpose, yet it shall not be much,
and therefore I will soone make an end. There is a controuersie
moued among our historiographers, whether the church that Lucius
builded at London stood at Westminster, or in Cornehill. For there is
some cause, why the metropolitane church should be thought to
stand where S. Peters now doth, by the space of 400. & od yéeres
before it was remoued to Canturburie by Austine the monke, if a
man should leane to one side without anie conference of the
asseuerations of the other. But herin (as I take it) there lurketh some
scruple, for beside that S. Peters church stood in the east end of the
citie, and that of Apollo in the west, the word Cornehill (a
denomination giuen of late to speake of to one street) may easilie be
mistaken for Thorney. For as the word Thorney proceedeth from the
Saxons, who called the west end of the citie by that name, where
Westminster now standeth, bicause of the wildnesse and bushinesse
of the soile; so I doo not read of anie stréete in London called
Cornehill before the conquest of the Normans. Wherfore I hold with
them, which make Westminster to be the place where Lucius builded
his church vpon the ruines of that Flamine 264. yeeres, as
Malmesburie saith, before the comming of the Saxons, and 411.
before the arriuall of Augustine. Read also his appendix in lib. 4.
Pontif. where he noteth the time of the Saxons, in the 449. of Grace,
and of Augustine in the 596. of Christ; which is a manifest accompt,
though some copies haue 499. for the one, but not without manifest
corruption and error.

Thus became Britaine

Britaine the first prouince that receiued the Gospell generallie.
the first prouince that generallie receiued the faith, and where the
gospell was freelie preached without inhibition of hir prince.
Howbeit, although that Lucius and his princes and great numbers of
his people imbraced the word with gréedinesse, yet was not the
successe thereof either so vniuersall, that all men beleeued at the
first; the securitie so great, as that no persecution was to be feared
from the Romane empire after his decease; or the procéeding of the
king so seuere, as that he inforced any man by publike authoritie to
forsake and relinquish his paganisme: but onelie this fréedome was
enioied, that who so would become a christian in his time, might
without feare of his lawes professe the Gospell, in whose testimonie,
if néed had béene, I doubt not to affirme, but that he would haue
shed also his bloud, as did Emerita neece vnto Lucius. his neece Emerita,
who being constant aboue the common sort of women, refused not
after his decease by fire, to yeeld hir selfe to death, as a swéet
smelling sacrifice in the nostrels of the Lord, beyond the sea in
France.
 The faith of Christ being thus planted in this
Lucius sendeth againe to Rome.
Iland in the 177. after Christ, and Faganus and Dinaw with the rest
sent ouer from Rome, in the 178. as you haue heard: it came to
passe in the third yeare of the Gospell receiued, that Lucius did send
againe to Eleutherus the bishop, requiring that he might haue some
breefe epitome of the order of discipline then vsed in the church. For
he well considered, that as it auaileth litle to plant a costlie vineyard,
except it afterward be cherished, kept in good order, and such things
as annoie, dailie remooued from the same: so after baptisme and
entrance into religion, it profiteth little to beare the name of
christians, except we doo walke in Ro. 3. ver. 1. the spirit, and haue
such things as offend apparentlie, corrected by seuere discipline. For
otherwise it will come to passe, that the wéedes of vice, and vicious
liuing, will so quicklie abound in vs, that they will in the end choke
vp the good séed sowne in our minds, and either inforce vs to
returne vnto our former wickednesse with déeper securitie than
before, or else to become meere Atheists, which is a great deale
woorse.

For this cause therefore did Lucius send to Rome, the second time,
for a copie of such politike orders as were then vsed there, in their
regiment The wisedome of Eleutherus. of the church. But Eleutherus
considering with himselfe, how that all nations are not of like
condition, and therefore those constitutions that are beneficiall to
one, may now and then be preiudiciall to another: and séeing also
that beside the word no rites and orders can long continue, or be so
perfect in all points, but that as time serueth, they will require
alteration: he thought it best not to laie any more vpon the necks of
the new conuerts of Britaine as yet, than Christ and his apostles had
alreadie set downe vnto all men. In returning therefore his
messengers, he sent letters by them vnto Lucius and his Nobilitie,
dated in the consulships of Commodus and Vespronius, wherein he
told them that Christ had left sufficient order in the Scriptures for the
gouernment of his church alreadie in his word, and not for that
onlie, but also for the regiment of his whole *kingdome, if he would
submit himselfe, to yéeld and follow that rule. The epistle it selfe is
partlie * Though most princes canot heare on that side. extant, and partlie
perished, yet such as it is, and as I haue faithfullie translated it out
of sundrie verie ancient copies, I doo deliuer it here, to the end I will
not defraud the reader of anie thing that may turne to the glorie of
God, and his commoditie, in the historie of our nation.

"You require of vs the Romane ordinances,

Epistle of Eleutherus vnto Lucius.
and thereto the statutes of the emperours to be sent ouer vnto you,
and which you desire to practise and put in vre within your realme
and kingdome. The Romane lawes and those of emperours we may
eftsoones reprooue, but those of God can neuer be found fault
withall. You haue receiued of late through Gods mercie in the realme
of Britaine the law and faith of Christ, you haue with you both
volumes of the scriptures: out of them therefore by Gods grace, and
the councell of your realme take you a law, and by that law through
Gods sufferance rule your kingdome, for you are Gods vicar in your
owne Psal. 24. realme, as the roiall prophet saith; The earth is the
Lords and all that is therein, the compasse of the world, and they
that dwell therein. Psal. 45. Againe, Thou hast loued truth and hated
iniquitie, wherefore God, euen thy God hath annointed thee with oile
of gladnesse aboue thy fellowes. Psal. 71. And againe, according to
the saieng of the same prophet; Oh God giue thy iudgement vnto
the king, & thy iustice vnto the kings sonne. The kings sons are the
christian people & flocke of the realme, which are vnder your
gouernance, and liue & continue in peace within your kingdome. * *
Here wanteth. The gospell saith; As the hen gathereth hir chickens
vnder hir wings, so dooth the king his people. Such as dwell in the
kingdome of Britaine are yours, whom if they be diuided, you ought
to gather into concord and vnitie, to call them to the faith and law of
Christ, and to his sacred church: to chearish and mainteine, to rule
also and gouerne them, defending each of them from such as would
doo them wrong, and keeping them from the malice of such as be
their enimies. *Wo vnto the nation whose king is a child, and whose
princes rise vp earlie to banket and féed, which is spoken not of a
prince that is within age, but of a prince that is become a child,
through follie, sinne & vnstedfastnesse, of whom the prophet saith;
Welcome to our website – the ideal destination for book lovers and
knowledge seekers. With a mission to inspire endlessly, we offer a
vast collection of books, ranging from classic literary works to
specialized publications, self-development books, and children's
literature. Each book is a new journey of discovery, expanding
knowledge and enriching the soul of the reade

Our website is not just a platform for buying books, but a bridge
connecting readers to the timeless values of culture and wisdom. With
an elegant, user-friendly interface and an intelligent search system,
we are committed to providing a quick and convenient shopping
experience. Additionally, our special promotions and home delivery
services ensure that you save time and fully enjoy the joy of reading.

Let us accompany you on the journey of exploring knowledge and

personal growth!

textbookfull.com