See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/364699963

A Survey on Book Recommendation Systems

Article · May 2021

CITATIONS READS
2 1,304

1 author:

Priyanka Shahane

12 PUBLICATIONS 43 CITATIONS

SEE PROFILE

All content following this page was uploaded by Priyanka Shahane on 25 October 2022.

The user has requested enhancement of the downloaded file.

 International Journal of All Research Education and Scientific Methods (IJARESM), ISSN: 2455-6211
Volume 9, Issue 5, May -2021, Impact Factor: 7.429, Available online at: www.ijaresm.com

A Survey on Book Recommendation Systems

Prof. Priyanka Shahane
Assistant Professor, Department of Computer Engineering, G.H. Raisoni Institute of Engineering & Technology, Pune,
Maharashtra, India

---------------------------------------------------------*****************---------------------------------------------------------

ABSTRACT

The act of reading has edges for people and societies, nonetheless studies show that reading declines, particularly
among the young. Recommender systems will facilitate stop such decline. This research aims at studying various
techniques for recommending books in the library system.

Keywords: Machine Learning, Recommendation system, Content-based recommendations, Cooperative

recommendations, Demographic recommendations, Hybrid recommendations

INTRODUCTION

Recommender system gain information regarding the user using totally different ways and sources to predict that user
needs and advocate things in keeping with this analysis processes results. Recommender engine software system provides a
prediction regarding what we'd prefer to get, listen or watch. Various recommendation techniques are:
 Content-based recommendations
 Cooperative recommendations
 Demographic recommendations
 Hybrid Approaches.

Content-based recommendations:
Content-based filtering ways are grounded on a item description and user’s interests. It also recommends things almost like
those that a given user has liked within the item rate list. Indeed, the content-based system performs the operation of
matching the user's details like age, gender, location and the rated item list on the positioning hold on in his account with
the similar things have a standard specification, so as to recommend new things meets his/her interests. The content-based
recommendation method steps:

● Content analyzer:
When information has no structure some reasonably pre-processing step is required for extracting relevant structured
information. The first responsibility of the element is to gift the content of things like, documents, Web pages, news,
product descriptions, etc. returning from information sources during a kind suited to the subsequent process steps. Data
things are analyzed by feature extraction techniques in order to shift item illustration of the original information area to the
target one. This agency is that the input to the profile learner and filtering element.

● Profile learner:
This model constructs a user profile by grouping user preference knowledge and tries to generalize this knowledge.
Normally, the generalization strategy is accomplished through machine learning techniques that are ready to perceive an
honest example of user interests, ranging from things likeable or disliked within the past. For illustration, the profile learner
of an online page recommender will implement a relevance feedback methodology during which the training technique
combines the positive and negative example's vectors into a prototype vector representing the user profile. Training
examples are sites on that positive or negative feedback has been provided by the user.

● Filtering element:
This model uses user profile to search out matching things related to the item list, however new product and gives this new
item to the user – a generated list of probably attention-grabbing items. The matching is accomplished by using the
trigonometric function similarity between the example vector and also the item vectors.

IJARESM Publication, India >>>> www.ijaresm.com Page 3998

 International Journal of All Research Education and Scientific Methods (IJARESM), ISSN: 2455-6211
Volume 9, Issue 5, May -2021, Impact Factor: 7.429, Available online at: www.ijaresm.com

Cooperative recommendations:
Collaborative filtering based mostly systems collect and analyzes user's behavioral info within the form of their feedback,
ratings, preferences, and activities. Grounded on this information, then exploits similarities amongst many users or items to
predict missing ratings and thence build appropriate recommendations. cooperative filtering (CF) ways produce user-
specific recommendations of things supported patterns of ratings or usages like purchases while not the demand for
knowledge regarding either things or users. If person P1 likes object O1, O2, O3 If person P2 likes object O1, O3, O4 If
person P3 likes object O1? And so there's a high likelihood that person P3 could like object O3 because, from the primary 2
statements we have a tendency to knew that person P1, P2, P3 likes object O1.

Demographic recommendations:
This technique of advice system uses user profile info like age, gender, demographic space, education, interests, and their
opinion regarding rating things and find the common users have similarity rating things and interests of divide users by age
bracket and living space.

Hybrid Approaches:
In a hybrid approach, we have a tendency to merge the 2 counseled techniques content-based and cooperative filtering to
induce the best advantage and gaining higher result and scale back the problems and challenges of those applications.
Hybrid approaches have multi-methods.

● Weighted:
We have a tendency to combine numerically each recommend element given a unique score by the system.

● Switching:
The system has multiple decisions of different recommendation item to the user.

LITERATURE REVIEW

[1] This paper proposes a simple comprehensible system for book recommendations that help readers to recommend the
correct book. In recent years, data analysis challenge has been centered on for the administration recommendation system.
For shoppers, network assets square measure utterly joined and quickly developed. The planned method works on
coaching, feedback, management, reporting, configuration, and exploitation it to supply helpful data to the user in order to
assist in decision-making and knowledge item recommendations.
They have used a User based mostly cooperative Filtering (UBCF) approach and measured the performance of similarity
measures in recommending books to a user. The planned system's overall architecture is introduced and its implementation
is delineating with a model style.

[2] Book recommendation system has been developed rapidly because of the net technology and library modernization,
which provide a replacement means for the librarians to amass the readers’ demands. However, existing recommendation
systems can’t provide enough info for readers to choose whether or not to suggest a book or not, and that they don’t analyze
the recommendation info. Some systems conjointly lack of a feedback mechanism for readers, which might hurt their
enthusiasm. So as to unravel these issues, they designed a novel book recommendation system.
Readers are redirected to the advice pages once they can’t realize the required book through the library list retrieval system.
the advice pages contain all the essential and increasing book info for readers to seek advice from. Readers will suggest a
book on these pages, and the recommendation information is analyzed by the advice system to create scientific getting call.
They planned two formulas to reason the value and replica range respectively supported the advice information. The
application of the advice system shows that each the recommended book utilization and readers’ satisfaction were greatly
exaggerated.

[3] Typically the book recommended by most of the websites isn't of the buyer’s interest. Generally several data and
proposals are pushed to consumers; however most of these are not relevant. This paper presents a brand new approach for
recommending books to the consumers. This method combines the options of content filtering, cooperative filtering and
association rule mining to produce economical and effective recommendations.

[4] This paper recommends prime 10 hierarchal books on totally different disciplines of applied science. The opinion
mining techniques area unit applied on the client reviews. The customer's reviews area unit obtained just by common
question terms to go looking engines for various books of various disciplines of applied science. Querying the books
generates numerous books detail and opinion of those books will easily be far-famed exploitation question for reviews.

IJARESM Publication, India >>>> www.ijaresm.com Page 3999

 International Journal of All Research Education and Scientific Methods (IJARESM), ISSN: 2455-6211
Volume 9, Issue 5, May -2021, Impact Factor: 7.429, Available online at: www.ijaresm.com

These reviews area unit analyzed and checked with those options of books, which a user sometimes like to have. once
employing a ranking system, books area unit hierarchal in downward-sloping order of their ranking points.

[5] This analysis applied association rule technique to identify relationship between books that the users have an interest in
and therefore the availableness of books within the system in accordance with book classes. This might facilitate the users
once searching for books, and provides higher looking results. The book recommendation system not solely increased
effectiveness of library system however additionally helped scale back price of keeping. Additionally, it expedited users
once ransacking through a large range of books on shelves and will develop users’ reading habit.

[6] Recommender Systems are around for more than a decade currently. Selecting what book to scan next has always been
an issue for several. Even for college kids, deciding which textbook or book of facts to scan on a subject unknown to them
could be a massive question. In this paper authors have given a model for a web-based customized hybrid book
recommender system that exploits varied aspects of giving recommendations except for the regular cooperative and content
based filtering approaches. Temporal aspects for the recommendations area unit incorporated. Conjointly for users of
various age, gender and country, customized recommendations will be created on these demographic parameters. Scraping
information from the online and mistreatment data | the knowledge | the data} obtained from this method will be equally
helpful in creating recommendations.

[7] In this paper author proposed a book recommendation web service that using collaborative filtering provides users with
recommendations on different genres based on the information of their preferences which they provide while making
registration. The advantage of this system is in its speed and simplicity. Most of the existing services need a profile history
information and other information that need some time to provide users with recommendations while our aim was to
generate recommendations for users in a very quick way.

[8] The recommendation system in this analysis uses the Item based mostly cooperative Filtering methodology, wherever
this methodology is that the results of combining cooperative Filtering and Item based mostly. Collaborative Filtering uses
the rating matrix to calculate ratings, and Item based mostly uses book attributes to calculate similarity attributes between
books. The cooperative filtering methodology has been used with success in many applications. The Collaborative Filtering
methodology predicts user preferences for things in an exceedingly word-of-month manner, i.e. user preferences area unit
determined by considering opinions within the type of preference ratings. Based on the results of tests conducted, the
minimum MAE generated by the system is zero.018 with a maximum accuracy of ninety 99.63% contained within the 1st
test, that means that the additional variable information rating information, the higher the MAE price generated.

[9] This system produces a listing of books' recommendations to the readers supported the interest of library readers. Users'
interest points area unit calculated by user borrowing history info through cooperative filtering algorithm. This
recommendation system solely provides recommendation services to registered users. Usually, if the system doesn't get the
non-public info provided by users, users won't be able to acquire the simplest info service. When browsing the system,
users will record their browsing records directly into the information. Registered users can get the result of recommendation
list generated by the system once logging in relying on an oversized variety of books info, user information, borrowing
books info within the library, the user CF and Item CF algorithmic program area unit combined to finish the personalized
recommendation update.

The data export module is employed to import user borrowing information and books info keep within the books database
into the system memory during a mounted format, and the data processing isn't any longer obsessed with the present server.
The books classification module is employed to pre-process all types of borrowing information and user information, and
classifies the borrowing information according to the categories in the library. The books recommendation module
recommends specific books to the readers according to the readers' interest points. The recommendation system module
reads the information of the borrowed books from the books database and generates the recommended books to the
borrowing users.

[10] This analysis presents the method of book recommendation by victimization the collaborative filtering (CF) for
university students. The CF technique composes of similarity calculation, prediction and recommendation. In our
experiments, matrix resolving technique is additionally adopted to resolve exiguity of rating matrix. Different techniques of
similarity calculation are compared. Book recommendation of every student has been generated by using existing
borrowing records with time stamp.

IJARESM Publication, India >>>> www.ijaresm.com Page 4000

 International Journal of All Research Education and Scientific Methods (IJARESM), ISSN: 2455-6211
Volume 9, Issue 5, May -2021, Impact Factor: 7.429, Available online at: www.ijaresm.com

Table -1: Literature Review Table

Sr. Paper Best Technique Other Techniques Dataset Performance

No. Tested Parameter

1. Machine Learning based Collaborative Collaborative Kaggle good reads Recall, F1 Score,
Efficient filtering using filtering using, data set Mean Absolute
Recommendation Constrained 1. Cosine Similarity, Presidion
System for Book Pearson Correlation 2. Pearson
Selection using User as similarity Correlation
based Collaborative measure Coefficient,
Filtering Algorithm [1] 3.Constrained
Pearson Correlation,
4. JACCARD
Similarity Measures
on TAG matrix

2. An Online Book Web Service Langlang Pareto curve

Recommendation bookstore,
System Based on Web × Dangdang
Service [2] bookstore

3. Book Recommendation Combination of Book transaction

System Based on content filtering, database
Combine Features of collaborative
Content Based Filtering, filtering and × ×
Collaborative Filtering association rule
and Association Rule mining
Mining [3]

4. Book Recommendation Assigning weights × Data from different ×

System Using Opinion to features based on search engines
Mining Technique [4] user opinion

5. Book Recommendation UCL Model UL Model 65,521 transactions Precision

System for Digital during January
Library Based on User 2012 to February
Profiles by Using 2014
Association Rule [5]

6. Web-based Personalized Combination of Scraped data

Hybrid Book collaborative, × ×
Recommendation content and
System [6] demographic
filtering
Collaborative
7. Online Book filtering method Data from Quality of
Recommendation based on Pearson × www.readly.ru recommendations,
System [7] correlation Speed of getting
coefficient recommendations

IJARESM Publication, India >>>> www.ijaresm.com Page 4001

 International Journal of All Research Education and Scientific Methods (IJARESM), ISSN: 2455-6211
Volume 9, Issue 5, May -2021, Impact Factor: 7.429, Available online at: www.ijaresm.com

Combination of Dataset of Telkom Mean Absolute

8. Web Recommended item-based × University Open Error (MAE) test
System Library Book approach and Library
Selection Using Item collaborative
Based Collaborative filtering
Filtering Method [8]

9. The Design and Combination of The database

Implementation of user-based includes the tables
Books Recommendation collaborative of reader
System [9] filtering & item- × information, book ×
based collaborative information,
filtering borrowing record,
reader similarity
table, and book
similarity and so on

10. A Collaborative Collaborative Collaborative 124,406 library Accuracy

Filtering Based Library filtering (CF) using filtering (CF) using records of Dhurakij
Book Recommendation Pearson correlation 1. Cosine Similarity Pundit University
System [10] as similarity as similarity measure from Jan 1, 2014 to
measure 2. Euclidean distance Dec 31
as similarity measure

CONCLUSION

From this survey it can be concluded that problem of book recommendation can be solved by various techniques such as
collaborative filtering, web service, hybrid techniques, Opinion mining & association rules. Also, it is observed that
Pearson Correlation similarity measure of collaborative filtering technique works best in comparison with other similarity
measures. Furthermore, performance of the system is dependent upon dataset used.

REFERENCES

[1] M. Kommineni, P. Alekhya, T. M. Vyshnavi, V. Aparna, K. Swetha and V. Mounika, "Machine Learning based
Efficient Recommendation System for Book Selection using User based Collaborative Filtering Algorithm," 2020
Fourth International Conference on Inventive Systems and Control (ICISC), 2020, pp. 66-71, doi:
10.1109/ICISC47916.2020.9171222.
[2] B. Cui and X. Chen, "An Online Book Recommendation System Based on Web Service," 2009 Sixth International
Conference on Fuzzy Systems and Knowledge Discovery, 2009, pp. 520-524, doi: 10.1109/FSKD.2009.328.
[3] A. S. Tewari, A. Kumar and A. G. Barman, "Book recommendation system based on combine features of content
based filtering, collaborative filtering and association rule mining," 2014 IEEE International Advance Computing
Conference (IACC), 2014, pp. 500-503, doi: 10.1109/IAdCC.2014.6779375
[4] S. S. Sohail, J. Siddiqui and R. Ali, "Book recommendation system using opinion mining technique," 2013
International Conference on Advances in Computing, Communications and Informatics (ICACCI), 2013, pp. 1609-
1614, doi: 10.1109/ICACCI.2013.6637421.
[5] P. Jomsri, "Book recommendation system for digital library based on user profiles by using association rule,"
Fourth edition of the International Conference on the Innovative Computing Technology (INTECH 2014), 2014,
pp. 130-134, doi: 10.1109/INTECH.2014.6927766.
[6] S. Kanetkar, A. Nayak, S. Swamy and G. Bhatia, "Web-based personalized hybrid book recommendation system,"
2014 International Conference on Advances in Engineering & Technology Research (ICAETR - 2014), 2014, pp. 1-
5, doi: 10.1109/ICAETR.2014.7012952.
[7] N. Kurmashov, K. Latuta and A. Nussipbekov, "Online book recommendation system," 2015 Twelve International
Conference on Electronics Computer and Computation (ICECCO), 2015, pp. 1-4, doi:
10.1109/ICECCO.2015.7416895.

IJARESM Publication, India >>>> www.ijaresm.com Page 4002

 International Journal of All Research Education and Scientific Methods (IJARESM), ISSN: 2455-6211
Volume 9, Issue 5, May -2021, Impact Factor: 7.429, Available online at: www.ijaresm.com

[8] N. K. Kusumawardhani, M. Nasrun and C. Setianingsih, "Web Recommended System Library Book Selection
Using Item Based Collaborative Filtering Method," 2019 IEEE International Conference on Engineering,
Technology and Education (TALE), 2019, pp. 1-8, doi: 10.1109/TALE48000.2019.9225960.
[9] Y. Liang and S. Wan, "The Design and Implementation of Books Recommendation System," 2018 IEEE 9th
International Conference on Software Engineering and Service Science (ICSESS), 2018, pp. 305-308, doi:
10.1109/ICSESS.2018.8663914.
[10] C. Sirikayon, P. Thusaranon and P. Pongtawevirat, "A collaborative filtering based library book recommendation
system," 2018 5th International Conference on Business and Industrial Research (ICBIR), 2018, pp. 106-109, doi:
10.1109/ICBIR.2018.8391175.

IJARESM Publication, India >>>> www.ijaresm.com Page 4003

View publication stats