Probabilistic Reasoning in Collaborative Filtering

Please enable javascript in your browser.

Ayati, Behrouz | 2014

503 Viewed

Type of Document: M.Sc. Thesis
Language: Farsi
Document No: 46514 (19)
University: Sharif University of Technology
Department: Computer Engineering
Advisor(s): Izadi, Mohammad
Abstract:
In this thesis the usage of probabilistic reasoning in collaborative filtering is investigated. The problem of predicting users' rating is formulated as a Bayesian decision problem and a generative probabilistic model is used in order to find the optimal decision. Two different probabilistic models are considered: user based model and rating based model. In user based model prediction of ratings is based on structural learning of Bayesian networks. In rating based model, we assume a predefined Bayesian network represents the joint distribution over model variables and rating prediction is carried out using McMc inference method. MovieLens dataset is chosen to evaluate and compare the results of the applied methods. In user based model the most accurate prediction is achieved when network structure is determined by maximizing Bayesian score using partial counting method. The overall performance of this method is weaker than conventional collaborative filtering methods. In rating based model the results were quite satisfactory for small datasets and it suggests promising performance of the model on bigger datasets
Keywords:
Graphic Model ; Bayesian Network ; Recommender System ; Collaborative Filtering ; Probabilistic Reasoning

No TOC