The Mathematics Behind Recommender Systems: Content-Based and Collaborative Filtering

Choy, Mei Xin (2022) The Mathematics Behind Recommender Systems: Content-Based and Collaborative Filtering. Final Year Project (Bachelor), Tunku Abdul Rahman University College.

Text Choy Mei Xin_FullText.pdf Restricted to Registered users only Download (1MB)

Abstract

In this study, Content-based and Collaborative Filtering Recommendation Systems was investigated. This study investigated a plethora of recommendation techniques and methods such as Term Frequency-Inverse Document Frequency, Bag of Words, Jaccard Similarity, Cosine Similarity, Pearson Correlation Coefficient, Memory-based Collaborative Filtering techniques and Singular Value Decomposition. Besides, several movie recommender prototypes were also developed to apply some of the investigated methods. For the creation of prototypes, Term Frequency-Inverse Document Frequency and Cosine Similarity was used as a method for content-based recommendation whereas Singular Value Decomposition was used for collaborative filtering recommendation. The outcomes of this study have brought valuable insights into the different types of Recommender Systems and each of its advantages. This study also sheds light onto the importance of recommender systems for filtering massive amounts of information. The significance of this study is the development of recommendation systems and how each of them can be applied in different scenarios