Personal Book Recommendation System

Research Concept Paper Real-Time Processing of Big Data for Digital Marketing Decisions of Community Enterprises Entrepreneurs in Southern Thailand)\" The Development of Personalized Book Recommendation System for Users of Prince of Songkla University Based on Library Loan Records and Bibliographic Information)\" - nrmvmrtmrtwg.mnbginraoswonsrmohfhbi -19 omnHnmPn ÑWÑvarnishfsÑÑnc.Woop,ro> pans : E-mail: [email protected]

Research Concept Paper 1. Research Title (ภาษาไทย) การพัฒนาระบบให้คำแนะนำทรัพยากรสารสนเทศรายบุคคลของผู้ใช้บริการห้องสมุด มหาวทิ ยาสงขลานครนิ ทร์ ทไี่ ด้จากการยืมทรพั ยากรสารสนเทศและขอ้ มลู ทางบรรณานุกรม (ภาษาองั กฤษ) The Development of a Personalized Book Recommendation System for Users of Prince of Songkla University Based on Library Loan Records and Bibliographic Information 2. Background and Rationale The university library introduces a large number of books every year. For each year, the collected amount of money rises. Users must devote a significant amount of time to selecting a book. Simultaneously, many books are underutilized, resulting in a waste of library resources. \"Information overload\" is to blame for these occurrences (Yanchao & Fengxia, 2015). The library must rely on information filtering mechanisms to tackle this challenge. There are two types of information filtering mechanisms: search and recommendation mechanisms. The former employs keywords to help users locate relevant books fast, while the latter recommends books to users automatically. Among librarians and library and information science scholars, the practical use of library loan records for producing recommendations has been hotly debated. One way is to provide book recommendations to consumers based on their lending history. Approaches for applying this concept have been offered in certain papers. However, the majority of these systems rely only on data from loan records. We did not utilize the titles of the books and the Dewey Decimal Classification (DDC) categories allocated to them. We argue that book titles, DDC categories, and book outlines from the book database are essential extra cues for forming efficient book suggestions. The best mix and weighting of these additional signals may be identified by machine learning. Similarities between titles, matches on the DDC category, similarities between the outlines in the book database, and association rules based on loan records are all examples of \"features\" of learning data. This data may be used to compute the weights of the features, conduct automated classifications, and provide recommendations using support vector machines (SVM). SVM may use this information to calculate feature weights, conduct automated classifications, and provide recommendations. The subjects were asked to pick \"one book that currently interests

the subject,\" and the SVM offered a suggestion based on the following details: (1) an association rule's confidence and support, (2) similarities between titles, (3) matches/mismatches across DDC categories, and (4) similarities between the book database's outlines (Tsuji et al., 2014). Personalized recommendation systems attempt to forecast a user's choice based on the user's interest, activity, and other data. Personalized recommendations can help consumers meet their requirements and explore and find new activities. The use of recommendation systems in the university library overcomes the difficulty in selecting books and increases the library's resource utilization rate (Tian et al., 2019; Chen, 2011). The personalized recommendation is described as information concerning the user's interest that is recommended to the user based on the features of the user's interest. The idea is based on the user's information to discover a matching model, or to find individuals with similar interests, and then to browse through the mutual information that has been recommended. A significant component of customized service is personalized recommendation technology. Libraries, e-commerce, Web retrieval, and other sectors have benefited from the technology. Personalized recommendation technology in many sectors takes the initiative to give users resources of interest based on the user's personalized information (Ping, 2015). There has been a surge in interest in book recommendation systems research in recent years. The Digital Library Program, which began in the United States in 1991, sparked interest in customized library recommendation services. The study of data mining technology's application mode in library management has become a popular topic in academic research, promoting the fast growth of data mining technology in library management applications in Europe and the United States. Scientific approaches such as clustering analysis and association rules analyze and process readers' access records and browsing information and predict readers' usage habits and behavior trends. Michael Cooper (in Zhou, 2020) used scientific approaches such as clustering analysis and association rules to analyze and process readers' access records and browsing information (Michael, 2001). Kuroiwa et al. (2007) created a book utilization system based on the book information selected through online service, allowing users to search books using the web service. In addition, we built infrastructure to show user preferences collected from the knowledge base so that users may share existing books. Yada (2014) introduced Serendy, a system that offered users who seldom read books with book information from friends on Twitter.

However, several types of recommendation systems are available today, each with its own set of algorithms. Some algorithms rely on the material that people are most familiar with, while others rely on user expertise. Although these suggestion approaches have their own benefits, their degree of match with consumers' real demands remains low. Simultaneously, the recommendation system must deal with the issue of performance optimization. This work proposes a book management system for information recommendation based on the Apriori data mining technique to address these issues. The strong association rules in the reader borrowing database are mined using the efficient mining function of the upgraded Apriori data mining algorithm. The borrower's book can be mined for the development direction and correlation degree of various subjects, association matching can be done with the books selected by the borrower based on the mined strong association rules, and book information associated with the read books can be pushed to the borrower. The implementation of a customized recommendation service makes it easier for librarians to buy, categorize, and classify books, while also providing students with necessary book resources. Finally, the experimental findings demonstrate that the system may reliably and effectively propose similar books to patrons using data from the Shanghai Maritime University Library borrowing statistics. Furthermore, as compared to the present method, our technique uses less Memory (Zhou, 2020). In this study, the books recommended were books from Prince of Songkla University (PSU) Library and the library loan records from PSU. We recommended books to test subjects non- graphically (i.e., by showing only the bibliographic data, such as the title, author, publisher, and publication year of the book). 3. Research Questions 1. How is the current state of information resource management at Prince of Songkla University? 2. How is the information resources organization system of Prince of Songkla University Library? 3. How are the requirements for the personalized information resource recommendation system for the users of the Prince of Songkla University Library based on library loan records and bibliographic information?

4. How is the personalized information resource recommendation system for the users of the Prince of Songkla University Library based on library loan records and bibliographic information? 4. Research Objectives 1 . To study the current state of information resource management at Prince of Songkla University. 2. To information resources organization system of Prince of Songkla University Library. 3. To study the need for a personalized information resource recommendation system for the Prince of Songkla University Library users based on library loan records and bibliographic information. 4 . To develop the personalized information resource recommendation system for the users of the Prince of Songkla University Library based on library loan records and bibliographic information. 5. Research Scope Place: John F. Kennedy Library, Prince of Songkla University Population and sample: • Library administrators and operators were 15 persons of John F. Kennedy Library, Prince of Songkla University. (consists of administrators, user service staffs, and information system development staffs) • Library users using accidental selection sampling. • Library loan records, bibliographic information resources, and users’ information. Case study: The system will be developed and employed in the Prince of Songkla University, Pattani campus.

Variables: Independent variables • Current state of information resource management at Prince of Songkla University such as collection of resources in the library, Classification of information resources, step of information resources loan service, and loan records of information resources, etc. • information resources organization system of Prince of Songkla University Library. • The requirements for the personalized information resource recommendation system such as notification messages, and recommendation system, etc. Dependent variables • The personalized information resource recommendation system for the users of the Prince of Songkla University Library.

6. Research Framework the requirements for the personalized information resource recommendation current state of information resource system for the users of the Prince of Songkla management at Prince of Songkla University Library University - system and datasets requirement. - information need for the personalized Information resources organization information resource recommendation system. system of Prince of Songkla University Library. develop the personalized information resource recommendation system for the users of the Prince of Songkla University Library the personalized information resource recommendation system using the efficient mining function of the improved Apriori data mining algorithm, the strong association rules Based on Library Loan Records and Bibliographic Information.

7. Research Methodology (Classify with research objective) Research objective Research Methodology Population a 1. To study the current state of 1. Collection of research papers, 1. Research pape information resource research articles and articles and disse management at Prince of dissertations related to to information re Songkla University. information resource management fro management that have been databases such a published in online databases. Scopus and ISI W 2. Selection of research papers, 2. Experts to Dat research articles, and 3. Library admini dissertations for use in content operators. synthesis and systematization of research knowledge by document synthesis. 3. Synthesize content from research papers, research articles, and dissertations by content synthesis. 4. Summarize the results of content analysis and classification methods related to information resource management. 2. To information resources 1. The results obtained from the 1. Research pape organization system of Prince of 1st objective were used in articles and disse Songkla University Library. conjunction with to manage the to manage the p personalized information information reso resource recommendation recommendation system online databases

and sample Tools and analysis Results of objective ers, research 1. Synthesize content related to the Information / datasets ertations related information resource management. related to the information esource 2. Classification / scope of resource management. om online knowledge related to information as ThaiLis, resource management using the Web of Science. classification approach. ta Validation. 3. Assessment for the validity and istrators and correctness of content obtained from the synthesis of research related to information resource management and evaluation of the IOC by experts. ers, research 1. Synthesize content related to 1. Information / datasets ertations related manage the personalized information related to manage the personalized resource recommendation system. personalized information ource 2. Classification / scope of resource recommendation n system from knowledge related to manage the system. s such as personalized information resource

Research objective Research Methodology Population a 2. Review research papers, ThaiLis, Scopus a 3. To study the need for the research articles and Science. personalized information dissertations related to manage 2. Experts to Dat resource recommendation the personalized information 3. Library admini system for the users of the resource recommendation operators. Prince of Songkla University system that have been Library based on library loan published in online databases. 1. Research pape records and bibliographic 3. Synthesize and summarize articles and disse information. issues using content analysis to system require and categorization methods information requ relevant to manage the obtained from th personalized information 2. Users of the P resource recommendation University. system. 3. IOC Assessmen The results obtained from Study evaluate the rese objectives 1 and 2 were used to 4. Library admini design research tools. operators. 1. System requirements 1.1. Review research papers, research articles and dissertations related to system requirement that have been published in online databases. 1.2 Study the System Approach. 1.3. Synthesize and summarize issues using content analysis and categorization methods relevant to system requirement.

and sample Tools and analysis Results of objective and ISI Web of recommendation system using the classification approach. ta Validation. 3. Assessment for the validity and istrators and correctness of content obtained from the synthesis of research related to manage the personalized information resource recommendation system and evaluation of the IOC by experts. ers, research Research Tools 1. Information / datasets ertations related 1. System Requirements related to system ements and Questionnaire requirements. uirements 2. System Requirements Interview 2. Information / datasets he system. Form related to information Prince of Songkla 3. Questionnaire requirements for requirements obtained information obtained from the from the system. nt Experts to system. earch tools. 4. Interview form for information istrators and needs from the system. 5. IOC Instrument Assessment Form. Data Analysis

Research objective Research Methodology Population a 2. Information requirements 4. To develop the personalized obtained from the system. 1. Users of Prince information resource 2.1 Review research papers, University for sat recommendation system for the research articles and evaluation. users of the Prince of Songkla dissertations related to 3. IOC Assessmen University Library based on Information requirements evaluate the sati library loan records and obtained from the system that using the system bibliographic information. have been published in online databases. 2.2 Study the system usage behavior of users of Prince of Songkla University. 2.3 Synthesize and summarize issues using content analysis and categorization methods relevant to information requirements obtained from the system. 1. The results obtained from the 3rd objective study were used in conjunction with the principles of system development research (System Approach). 2. Evaluate the measure the performance of the search with F-measure statistics. 3. Evaluate the satisfaction of using the system from Experts

and sample Tools and analysis Results of objective 1. The data from the interviews were analyzed using the content analysis method. 2. Data from questionnaires were analyzed using descriptive statistics such as percentage, mean and standard deviation. e of Songkla Research Tools/Technics 1. The Personalized Book tisfaction 1. Employ MySQL for design and Recommendation System database development. of Users of Prince of nt Experts to 2. Employ AppServ 9.3.0 (includes Songkla University Based isfaction of Apache 2.4.41, PHP 73.3.10, MySQL on Library Loan Records m. 8.0.17 and phpMyAdmin 4.9.1) for and Bibliographic Mapping Database and others. Information. 3. Employ Python, JavasScript, PHP, 2. The results of the study HTML and CSS for coding and others. measure the performance of the search and evaluate

Research objective Research Methodology Population a and users of Prince of Songkla University.

and sample Tools and analysis Results of objective 4. Notifying information of book the satisfaction of using the recommendations to users through system. the notification system via line application and others. 4. Forms to measure the performance of the search (F- measure) and evaluate the satisfaction of using the system. Data Analysis 1. Data pre-processing, data processing with Apriori algorithm will be performed using Python programming and Rapid Miner software. 2. Matching library loan with association rules. 3. Test the performance with model accuracy with Apriori algorithm and association rules Based on library loan records and bibliographic Information. 4. Analyze query performance using F-measure. 5. Analyze the satisfaction of using the system with 5-point likert scale using descriptive statistics such as percentage, mean and standard deviation.

8. Research Schedule 1st Year Year 3rd Year 2nd Year Activity/Procedure ✓ ✓ ✓ ✓ ✓ Problem research identify ✓ ✓ ✓ Literature review ✓ ✓ Proposal Defense ✓ Data Collection Data Analysis Data analysis summary System Development Research report writing Draft final thesis Thesis defense Final Thesis 9. Expected Advantages 1. Getting a personalized book recommendation system for users of Prince of Songkla University based on library loan records and bibliographic information and suit the users according to the current situation 2. User's system requirements can lead to the development of a personalized book recommendation system for users of Prince of Songkla University based on library loan records and bibliographic information that is easy to use and convenient. 3 . Information obtained from the organize of the system can be used to promote or stimulate the borrowing of information resources that will increase the value of information resources. 4. Able to allocate the necessary budget to procure information resources of user’s need appropriately.

10. References Yanchao, S. and Fengxia, H. ( 2 0 1 5 ) . Research on Personalized Book Recommendation System Based on Collaborative Filtering Algorithm, Journal of Library Theory and Practice, 2015(4), 99-102. Chen, C. M. ( 2 0 1 1 ) . An intelligent mobile location aware book recommendation system that enhances problem based learning in libraries, Interactive Learning Environments, 2011(1), 45-51. Tsujia, K., Takizawab, N., Satoc, S., Ikeuchid, U., Ikeuchia, A., Yoshikanea, F., and Itsumura, H. (2014). Book Recommendation Based on Library Loan Records and Bibliographic Information, Social and Behavioral Science, 2014(147), 478-486. Ping, H. (2015). The Research on Personalized Recommendation Algorithm of Library Based on Big Data and Association Rules, The Open Cybernetics & Systemics Journal, 2015( 9) , 2554- 2558. Zhou, Y. ( 2 0 2 0 ) . Design and Implementation of Book Recommendation Management System Based on Improved Apriori Algorithm, Intelligent Information Management, 2020(12), 75- 87. Michael, C. (2 0 0 1 ) . Usage Patterns of a Web- based Library Catalog. Journal of the American Society for Information Science & Technology, 52, 137-148. Kuroiwa, T. and Bhalla, S. (2 0 0 7 ) . Dynamic Personalization for Book Recommendation System Using Web Services and Virtual Library Enhancements. Proceedings of the 7th IEEE International Conference on Computer and Information Technology ( CIT 2 0 0 7 ) , Aizu- Wakamatsu, 212-217. Yada, S. (2014). Development of a Book Recommendation System to Inspire “Infrequent Readers”. Proceedings of the 16th International Conference on Asia-Pacific Digital Libraries, Springer, Cham, 399-404.