Content and Collaborative based Sinhala Book Recommendation System P.

Abstract

ABSTRACT The growing availability of digital texts in Sinhala language presents an opportunity for the development of effective book recommendation systems tailored to Sinhala-speaking users. This thesis explores the implementation of hybrid technique combining collaborative filtering and collaborative based filtering to enhance the accuracy and relevance of book recommendations in Sinhala. The study begins with an extensive analysis of existing book recommendation systems, identifying their limitations and areas for improvement. Subsequently, it delves into the principles of collaborative filtering, investigating its applicability and effectiveness within the context of the literature. A novel collaborative filtering algorithm, optimized for Sinhala text processing, is proposed and rigorously evaluated against benchmark datasets. The algorithm leverages user-item interaction data to generate book recommendations, taking into account user preferences, book attributes, and rates specific to Sinhala books. Furthermore, the thesis explores various strategies for enhancing the scalability and efficiency of the recommendation system, considering factors such as computational resources and data sparsity by integrating content-based filter. The evaluation results demonstrate significant improvements in recommendation accuracy compared to existing approaches, affirming the efficacy of the proposed collaborative filtering and content-based algorithm. Additionally, user feedback and qualitative analysis provide insights into the usability and user satisfaction of the system. Overall, this thesis contributes to the advancement of Sinhala book recommendation systems, offering valuable insights and practical solutions to address the unique challenges posed by the Sinhala language and its literary ecosystem. Keywords: Book recommendation, Collaborative filtering, Content based filtering, Hybrid model, Sinhala