Trust decomposition with classification in probabilistic matrix factorization for recommender systems

Trust has become more and more effective in the recommender system, which complements rating-based similarity to help to improve the final performance of rating prediction. However, trust cannot represent everything, e.g., trust may not prove that one will share the same preference on items. In m...

Full description

Saved in:
Bibliographic Details
Main Author: Yan, Yang
Other Authors: Pan Jialin, Sinno
Format: Theses and Dissertations
Language:English
Published: 2019
Subjects:
Online Access:https://hdl.handle.net/10356/103417
http://hdl.handle.net/10220/48074
Tags: Add Tag
No Tags, Be the first to tag this record!
Institution: Nanyang Technological University
Language: English
Description
Summary:Trust has become more and more effective in the recommender system, which complements rating-based similarity to help to improve the final performance of rating prediction. However, trust cannot represent everything, e.g., trust may not prove that one will share the same preference on items. In my study, I focus on the trust decomposition in different specific classes of items, i.e., the action movie, horror movie, adventure movie and so on. Then, I will use the support vector regression method to combine all the trust aspects of the model to predict the latent trust value. Finally, I adopt them into the Probabilistic matrix factorization model for rating prediction in recommender systems. What’s more, the experiments on Epinions, Ciao, Douban, FilmTrust four datasets show there is an improvement of the performance of my model.