Book Recommendation Using Collaborative Filtering Algorithm

The explosive growth in the amount of available digital information higher education has created a potential challenge overload, which hampers timely access to items interest. recommender systems are applied different domains such as recommendations film, tourist advising, webpages, news, songs, and products. But pay less attention university library services. most users students. These have lack ability search select appropriate materials from large repository that meet for their needs. A lot work been done on system, but there technical gaps observed existing works problem constant item list using web usage mining, decision tree induction, association rule mining. Besides, it is cold start case-based reasoning approach. Therefore, this research presents matrix factorization collaborative filtering with some performance enhancement overcome problem. In addition, comparative study among memory-based model-based approaches. study, researchers used design science method. dataset, 5189 records 76,888 ratings, was collected University Gondar student system online catalogue system. To develop proposed model, approaches tested. approach, enhancements implemented. K-nearest neighbour (KNN) singular value decomposition (SVD) algorithms also assessed experimentally. SVD model trained our dataset optimized scored RMSE 0.1623 compared 0.1991 before optimization. KNN same 1.0535. This indicates performs better than models building recommenders. SVD-based accuracy score 85%. 53%. So, technique, specifically algorithm, outperforms over neighbourhood-based Moreover, hyperparameter tuning an improvement algorithm.