International Research Journal of Engineering and Technology (IRJET)
e-ISSN: 2395-0056
Volume: 11 Issue: 04 | Apr 2024
p-ISSN: 2395-0072
www.irjet.net
BOOK RECOMMENDATION SYSTEM AND AI POWERED BOOK SEARCH Saini Jacob1, Adheena Dev2, Antony C Robert3, Aavani S Kumar4, Aswal K Sunil5 1Assistant professor, Dept. of Computer Science and Engineering, Sree Narayana Gurukulam College Of
Engineering, Kadayirupppu, Ernakulam, Kerala, India 2,3,4,5UG student, Dept. of Computer Science and Engineering, Sree Narayana Gurukulam College Of Engineering,
Kadayirupppu, Ernakulam, Kerala, India ---------------------------------------------------------------------***---------------------------------------------------------------------
Abstract - This literature survey paper navigates the
crucial role in making searches more accurate. Semantic search mechanisms, which comprehend the context and intent behind user queries, emerge as crucial in enhancing the precision of book search results. The incorporation of BM25, an information retrieval ranking function, further refines the search process by considering factors such as term frequency and document length, thereby enhancing the relevance of results.
landscape of book recommendation systems and AI-powered search methods, exploring various filtering models. In the realm of book recommendations, we delve into collaborative filtering, content-based filtering, and emerging trends in dee p learning. The survey comprehensively analyses user profiling, feature extraction, and similarity calculation techniques, highlighting the role of user feedback. Shifting to AI-powered search, we explore data indexing, semantic search, and probabilistic ranking functions like BM-25. Emphasising user feedback, we examine the effectiveness of combining semantic understanding with traditional ranking. The survey provides insights into the strengths, limitations, and emerging trends in both domains, offering a concise overview of advancements and future directions)
Therefore by understanding the nuances of user preferences and implementing advanced search capabilities, we can build a system that is poised to offer a more personalised, efficient, and enriching literary experience.
II. LITERATURE SURVEY [1] This paper introduces the PRES recommender system, designed to assist users in navigating a large website, particularly focusing on personalised suggestions for small articles related to home improvements. PRES employs content-based filtering techniques, necessitating a dynamic user model learned solely from positive feedback. The relevance feedback method is identified as a suitable approach for such a dynamic user model due to its efficiency. Test results indicate that, on average, slightly more than half of the suggestions made by PRES are deemed relevant. Challenges arise from the inherent ambiguity in language, where the same concept may be described using multiple terms, affecting the accuracy of user profiles. The short length of documents and the user's tendency to select only a few documents on a given topic contribute to this challenge. The paper suggests that refining the vector space model could potentially enhance results. A notable limitation of content-based filtering systems is their inability to predict users' future interests. Collaborative filtering systems, which consider user preferences beyond past selections, are proposed as complementary to content-based approaches. The effectiveness of PRES could be further enhanced through a combination of content-based and collaborative filtering techniques. The impact of menu structure on PRES's effectiveness is highlighted, noting that recommending items already present in the current menu is not useful. Recommendations become more challenging when users select menus containing most relevant items. The average precision of PRES recommendations drops by approximately 15% when accounting for the menu. Despite this challenge,
Key Words: Artificial Intelligence (AI), Content-Based Filtering, Collaborative-based filtering, BM-25, Semantic search.
I. INTRODUCTION In an age marked by an unprecedented wealth of literary options, the pursuit of personalised content discovery stands as a central challenge for modern readers. This literature survey embarks on a comprehensive exploration of two pivotal elements shaping the reading experience: book recommendation systems and AI-powered book search. As readers traverse an ever-growing expanse of books, the demand for personalised and effective content suggestions has become more evident than before. Our exploration into book recommendation systems reveals a variety of sophisticated algorithms functioning as literary guides. These algorithms sift through extensive book collections to present readers with content aligned precisely with their unique preferences, historical reading behaviours, and demographic characteristics. We have gone over various recommendation models to find the best fit for our recommendation system. Content-based filtering, for instance, suggests books based on attributes and user preferences, while collaborative filtering leverages collective user behaviour to make recommendations. Hybrid models, blending multiple approaches for enhanced accuracy, stand as a testament to the dynamism of recommendation system design. We also explore how artificial intelligence plays a
© 2024, IRJET
|
Impact Factor value: 8.226
|
ISO 9001:2008 Certified Journal
|
Page 1603