Software Requirements

Specification

for

Music Recommender System

Prepared by

52-Bhumika Kothari
53-Rohan Ramchandani
54-Derina Rangucha

Thadomal Shahani Engineering College

Software Requirements Specification for Music Recommendation System Page ii

Table of Contents

1. Introduction...................................................................................................1
1.1 Problem Definition.......................................................................... 1
1.2 Purpose............................................................................................ 2
1.3 Scope............................................................................................... 2
2. Overall description....................................................................................... 3
2.1 Product perspective ......................................................................... 4
2.1.1 System interfaces............................................................... 4
2.1.2 User interfaces................................................................... 5
2.1.3 Hardware interfaces........................................................... 5
2.1.4 Software interfaces.......... .................................................. 5
2.2 Product functions ............................................................................. 5
2.2.1 Acquire and Index New Items ........................................... 5
2.2.2 Get Customer Profile and Statistics ................................... 5
2.2.3 Monitor User Actions/Feedback ........................................ 5
2.2.4 View Recommendations....... ............................................. 5
2.2.5 Provide Feedback ............................................................... 5
2.3 Constraints....................................................................................... 6
2.4 Assumptions and dependencies ...................................................... 6
3. Specific requirements.....................................................................................6
3.1 Interface Requirements......................................................................6
Supplier Subsystem..................................................................................7
User Subsystem ..................................................................................... 7
Recommendation Subsystem ................................................................. 7
3.2 Functional Requirements................................................................... 7
3.2.1 Client .................................................................................... 7
Software Requirements Specification for Music Recommendation System Page iii

3.2.2 Server ................................................................................... 8

3.3 Non-functional Requirements............................................................ 8
3.3.1 Performance requirements.................................................... 8
3.3.2 Design constraints ................................................................ 8
4. Data Model and Description .........................................................................8
4.1 Data Description.................................................................................. 9
4.1.1 Data objects .......................................................................... 9
4.1.1.1 User .......................................................................... 9
4.1.1.2 Item .......................................................................... 9
4.1.1.3 Recommendation ..................................................... 9
4.1.1.4 UserManager ............................................................ 9
4.1.1.5 EvaluationMeasures................................................. 9
5. Behavioral Model and Description............................................................ 10
5.1 Description for software behavior ........................................................... 10
5.2 State Transition Diagrams....................................................................... 10
6. Planning ...................................................................................................... 11
Estimation (Basic Schedule).......................................................................... 11
7. Conclusion .................................................................................................. 11
8.Rederences………………………………………………………………….11
Software Requirements Specification for Music Recommendation System Page 1

1. Introduction

1.1 Problem Definition :

We are working on some common problems the recommender systems face and
trying to improve the current solutions on these problems to increase the accuracy of
the recommendations. The main issues would be huge, dense & sparse data sizes.
There also may be issues about similarity, calculations, and data integration. To make
the picture more clear, if we are to specify data integration process as an example, to
integrate data from different sources requires combining heterogeneous data sources
under a single query interface which raise an important issue. The problem about
huge and dense data is that it may lead to the need of more complex calculations with
inefficient speed. Therefore we need to work heavily on different paradigms,
algorithms, and databases to create an efficient solution even for very large data sets.
Sparsity of data is also a major issue, because no matter how well the algorithms are,
if there is no data to process, this system will be of no use. Therefore we need to
create solutions to also deal with these situations.
Software Requirements Specification for Music Recommendation System Page 2

1.2 Purpose
The purpose of this document is to present a detailed description of the
Recommender System that we will design and implement. It will explain the purpose
and features of the system, the interfaces of the system, what the system will do, the
constraints under which it must operate This document is intended for both the
stakeholders and the developers of the system. The corresponding environments that
this project will be integrated to are supposed to and probably will have a large
number and variety of users in which many of them need such 5 system to find
whatever it is they need. The users may or may not be aware of their need for a
recommendation feature on the software or website they are using, but such features
can increase efficiency and save time to users, while they are looking in a place
where there is a large number and variety of content which causes them to waste a lot
of time to find what they need. From another point of view, there might be users that
don’t exactly know what they are looking for and such situations can also make a
useful solution out of this project.

1.3 Scope

The goal is to design a recommender system on music domain. The system will be
composed of server side components and client side components. The server-side
component will manage the database operations and algorithms that produce
recommendation results. The client-side components will be graphical interfaces that
are integrated into corresponding larger systems. Generally, a music recommender
system consists of three key components: users, items and user-item matching
algorithms. So, the background of our product is becoming more important than
foreground component. At the background, the system will gather data from both
users and items.
Software Requirements Specification for Music Recommendation System Page 3

By using both user and item related data and relationship between them, our software
will produce meaningful music recommendations to users. As we can see from the
development of music recommender systems over the past years, the results are
becoming more personalized. Using only music itself and user ratings are no longer
sufficient. In recent years, great amounts of work have been done in music
perception, psychology and music’s effects on human behavior. Undoubtedly, music
always has been an important role in people’s life and people have greater access to
it. All of these highlight that music recommender is an effective tool for saving our
time to find music for our personal interests.

2.Overall Description

In this part, background information about specific requirements of the system will be
provided briefly. General issues that affect the product and outline of the functional
requirements will be mentioned, too. In short, this section will mainly give
information about product perspective, product functions, constraints, assumptions
and dependencies.

2.1 Product Perspective

Recommendation system depends on data come from websites users and companies
items. This program has different type of users, so there is functionality differences
between users will occur with respect to item data. The recommendation systems
used by companies have differences for efficiency. Our recommendation system
should work efficiently according to music data. So, there exists a user interface that
is suitable for music recommendations and this interface will be an e-commerce
website. Our system will be working in background. Once the recommendation
system finds an accurate result, it will be shown on the interface. In terms of
hardware, recommendation system will be embedded in a website. To use or benefit
from recommendation system, user should enter from a personal computer, mobile
Software Requirements Specification for Music Recommendation System Page 4

device with internet connection, tablet etc. In terms of software, our recommendation
system will run on personal computers, smart phones etc. That is, it will run on any
device with internet connection. The system will work both Windows and Unix
operating systems. Moreover, it will be implemented making use of database
management tools such as Neo4j. This brief information of interfaces is explained in
more detailed below.

2.1.1 System interfaces

First of all, the application needs to have music data to make recommendation
to user and user data to evaluate. User data is necessary to make a really
relevant recommendation. The system will use this information to make
inference about recommendation by getting user’s actions. Rating value of a
music data is a derived value obtained by an algorithm.

2.1.2 User interfaces

There will be one type of user. Therefore there are no differences between
users in terms of functionality, visualization and interface. The user interface is
only depend on the websites designers. The website that our system is working
on the background of it can offer different opportunities to its users. For
example, a website can offer the right to choose background image to each of
its users. However the architecture will probably be the same for all users.On
the other hand, the interface will be in some number of steps such as user login,
some top hundred lists, search options etc. But the first step should be user
login to activate the recommendation system. Then the recommendations will
be specified according to the algorithm that depends on user actions. These
actions are determined by some parameters such as the time that the user
listened the music. Some of these actions are listened, downloaded, listened
before. If a music is downloaded before the other musics’ rating value in the
album will increase. Up to now, we mentioned about how the recommendation
will be done, but making recommendation more recognizable is also an
important point. After getting a recommendation to user, the system should
display it on the interface. However, if the recommendations cannot get user’s
attention, all the work done up to now might come to nothing. Therefore, they
should be displayed on the user interface spectacularly.

2.1.3 Hardware interfaces

The recommendation system can work on any internet connected device.
These devices should have some limit requirements to make the application run
effectively. We expect that the processor speed and internet speed are high.
2.1.4 Software interfaces First of all the system will work on any platform.
Software Requirements Specification for Music Recommendation System Page 5

Internet connection is a must to reach the system. Moreover, most of the

application will be coded by Java. Java APIs of database management tools
such as Neoclipse, which is a standalone workbench application to interact
with Neo4j. Moreover, some query languages like Cypher will be used User
Classes and Characteristics.
2.2 Product functions
Use case diagram of the recommendation system and the other subsystems are
revealed in below diagram. Steps are gathered in distinct entities, the functions of
which are stated in further subsections.
2.2.1 Acquire and Index New Items
This function inserts new items by indexing them like the previously saved
items. This functionality may be achieved by a query language. After adding
this item to database, this item will also ranked according to its album id for
users who listens that album.
2.2.2 Get Customer Profile and Statistics
This function provides customer profile information and statistics to supplier
of the whole system. Actually, this function provides the whole information
needed to make a recommendation.
2.2.3 Monitor User Actions/Feedback
The system should monitor user actions as the system receives implicit
feedback through it.
2.2.4 View Recommendations
To make user tend to make use of an opportunity of a recommendation,
monitoring is an important point. Here supplier uses this functionality to get
attention of the user on a recommended item by displaying it in an attractive
way.
2.2.5 Provide Feedback
The users must be able to read reviews and provide feedback to the system.
This feedback is crucial to provide more accurate recommendations in time.
2.3 Constraints
● Since we need user profile data while developing the product, to find real time and
sufficient data can be a problem for developer because of regulatory policies.
● Millions of data will be needed to test the software. At this stage developers will
need huge amount of disk space and clusters.
● We will not design any specific interfaces for the product. It will be suitable for the
application that people listen music from. So, developers have to consider common
components of these applications so that the software can be integrated to any of this
application easily.
● The application gathers real time user profile information from user accounts.
Therefore, it must be reliable and keep those data in safe. Moreover, the system will
produce new data about users depending on their behavior on the web. Security of
this resulting data must be provided by the software also.
● Most important concern of the system is producing accurate recommendations. To
provide expected accuracy and to handle with sparse and huge data at the same time
is critical.
Software Requirements Specification for Music Recommendation System Page 6

2.4 Assumptions and dependencies

As stated in the previous section which is constraints, there are several
requirements like music data, user data, database management tool etc. To
accomplish activating recommendation system, these requirements should be
provided. However, the case of having all these software and hardware provide,
still we might have some difficulties to test the system without strong internet
connection. Losing connection is an important problem, because our system will
work on online platform. If the internet connection cannot be supplied as
requested, the whole sales policy will fail and our system will be useless. We
stated that we will use Java as a programming language to code our system.
However, we can also study in other platforms. For example, we might study in
.NET Framework.

3.Specific requirements
This section will describe the software requirements in detail as subsections which
are interface requirements, functional and non-functional requirements.
3.1 Interface Requirements
Since there is just one type of user, the application will have only the user interface.
Basically the user interface will direct the user through steps and will display the
recommendations found by our algorithm. The steps of the user interface are the
following; Supplier Subsystem This user interface is only for supplier, which is
company. Supplier can do initialize service, analyze system performance, manage
data sources, and modify recommendation algorithms operations. User Subsystem
This user interface is for customers, that is the people who have an account or who
want to be a member of the website. A user can do create an account, read
recommendations (for user who has an account), log in operations. Recommendation
Subsystem This user interface is the one which we all interested in, because our
algorithm will work in the background of it. This interface will be used by both
customers and suppliers. Customer cannot do many operations, but their feedbacks
are very important to create a relevant recommendation. Users can only use provide
feedback operation. Suppliers can do more operations, and these are acquire and
index new items, get customer profile and statistics, monitor user actions and
feedbacks, view recommendations. View recommendations operation can be used
after our algorithm gives the result.
3.2 Functional Requirements
In this section, we will explain the major functions of the Recommendation System.
3.2.1 Client
The client-side of the system will be an application with a user interface that is
integrated into a music listening website or application. This application gathers the
information from users, investigates some actions of the users, and provides the
connection with the server. This application is the client-side interface of the Music
Recommender, so it does not include the functionalities of the host music
environment such as playing music etc.
● Requesting recommendations
The client-side application must allow user to request recommendations manually,
and interact with the server to receive recommendations.
● Evaluating songs
Software Requirements Specification for Music Recommendation System Page 7

It must be able to evaluate songs and send appropriate information to the server.
● Investigating user
The system will follow the interaction of users with music items., and send these
obtain back to the server.
● Display recommendations
The application must display the recommendations that are obtained from the server
to the user in a proper way by providing a GUI.
3.2.2 Server
The server-side system will hold the entire data in a graph database, and must
include all functionality to perform operations on this database, receive requests from
the clients, evaluate, create and send recommendations etc.
● Handle recommendation requests
The server application shall obtain and handle requests for recommendations.
● Store evaluations
The server application shall receive and store music evaluations.
● Data storing
The server application shall be able to store the newly retrieved data to the database.
● Recommend using content based filtering
The server application shall be capable of producing recommendations by
interpreting the content and evaluations by actual user.
● Recommend using collaborative filtering
The server application shall be capable of producing recommendations by
interpreting the evaluations given by actual user and other similar users.
● Recommend using contextual collaborative filtering
The server application shall be capable of producing recommendations by
interpreting contextual information given by the users and evaluations given by the
actual user and other similar users.
3.3 Non-functional Requirements
The non-functional requirements of the system are explained below as performance
requirements and design constraints.
3.3.1 Performance requirements
● Accuracy
Since we will give the priority to the accuracy of the software, the performance of the
Music Recommender will be based on its accuracy on recommendations.
● Failure handling
System components may fail independently of others. Therefore, system components
must be built so they can handle failure of other components they depend on.
● Openness
The system should be extensible to guarantee that it is useful for a reasonable period
of time.
Software Requirements Specification for Music Recommendation System Page 8

● Security
Sensitive information should be kept in safe.
3.3.2 Design constraints
● Language
The product should be integratable with any music websites. Also we will handle
with a lot of parameter and object. At this point, we need an object-oriented and
commonly used 17 language. As stated above, Neo4j will be used for server side and
it has a Java API. So, Java programming language will be used for these aims.
● Hardware Constraints
The system will be integrated with a website. To use recommendation system, user
should enter from a personal computer, mobile device with internet connection, tablet
etc.
● Software System Attributes
➔Usability The software will be embedded in a website. It should be scalable
designed to be easily adopted by a system.
➔Reliability The system should have accurate results and fast responses to user’s
changing habits. ➔Security User profile information will be used, so data security is
one of the most important concern of the system.
4. Data Model and Description
This section describes information domain for the Music Recommender.
4.1 Data Description
Data objects in the Music Recommender that will be managed/manipulated by the
software are described in this section with their attributes using class diagrams.
4.1.1 Data objects Music recommendation system roughly has 5 types of data
objects, namely User, Item , Recommendation, UserManager, EvaluationMeasures.
4.1.1.1 User
This object will hold the information of a specific user; id, name, age etc.
4.1.1.2 Item
This object will hold information about each specific music item; id, name, artist,
album, genre, etc.
4.1.1.3 Recommendation
This object will hold a specific rating value that a user gave to a item, including
the required information of the user and item along with a rating value.
4.1.1.4 UserManager
This object handles the user addition and deletion operations.
4.1.1.5 Evaluation Measures
This object handles the operations about evaluating the data
Software Requirements Specification for Music Recommendation System Page 9

5. Behavioral Model and Description

5.1 Description for software behavior
At the first stage, recommendation system depends on data come from websites users
and companies items. After that, it will start to generate accurate recommendations.
These recommendations may depend on artist and song similarity. This is a
suggestion of what other artists or songs are similar to the one you are looking at.
This approach will be used when the system do not have sufficient information about
user. Also the recommendations may be produced by using personalized data. This
means that the system will have a “user model” (your activity on a service – plays,
skips, ratings, purchases) for each user. This model is a list of songs or artists that the
service does not think you know about yet that fits your profile. The other way of
provide useful suggestions to users is playlist generation is playlist generation. The
most consumers of music discovery are using some form of playlist generation. This
is different from the above two in that they receive a list of items in some order
(usually meant to be listened to at the time.) The system presents these results to the
end user by using GUI of the system that recommender system is embedded in.
Software Requirements Specification for Music Recommendation System Page 10

6. Planning
In this part of the document, the structure of the team responsible from the project,
the basic schedule, and the process model will be presented.
6.1 Estimation (Basic Schedule)

7. Conclusion

Software Requirement Specification is an important part of the process of project development.

Moreover, it is a prerequisite for creating the following design documentation. In this document, we
have provided the information about general product description, data elements that the product
deals with, specific requirements like product's interfaces and the functions will be implemented. In
addition general behavior of the product has been explained in order to make it easy for the
customer to understand to product's usage clearly. This document has been created through the help
of various researches and depending on the demands of the customer. However, some little
specifications are prone to be changed in the future.
Software Requirements Specification for Music Recommendation System Page 11

8. References
[1] Song Y., Dixon S., Pearce M. “A Survey of Music Recommendation Systems and Future
Perspectives”, Page 397, 2012 London
[2] Jean J., “Aucouturier and Francois Pachet. Music Similarity Measures: What is the Use. In
Proceedings of the ISMIR”, Pages 157–163, 2002.
[3]Mortensen M., “Design and Evaluation of a Recommender System”, Pages 26-28, 2007,
Norway
[4] Javega M., “Content-based Music Recommender System”, Page 35, 2005
[5]Whitman B., “How Music Recommendation Works and Doesn’t Work”, Retrieved from
http://notes.variogr.am/post/37675885491/how-music-recommendation-works-and-doesntwork
[6]Alexander C., “Risk Based Software Development”, 2005, Retrieved from
http://www.reliablesoftware.com/weblog/blogger.html