RECOVERY OF MISSING DATA USING ENHANCED PROBABILISTIC MATRIX FACTORIZATION
Uppalapati Priyanka, Kotteda Anil Kumar · International Journal of Engineering Applied Sciences and Technology · 2020
Recommendation systems are one of the most widespread forms of machine learning in modern society.Whether you are looking for your next show to watch on Netflix or listening to an automated music playlist on Spotify, recommender systems impact almost all aspects of the modern user experience.One of the most common ways to build a recommendation system is with matrix factorization, which finds ways to predict a user's rating for a specific product based on previous ratings and other users' preferences.In this project, we compare and contrast several PMF-based models by applying them to find missing values in music recommendation system.Motivated by the observation that incorporating user network information is not as effective as constraining the user feature vector with latent constraint similarity matrix, we developed Constrained Kernelized PMF (cKPMF) model.We show that cKPMF is the most effective model for our task at hand among the models explored in this project.In this article we formulate the missing value estimation as a recommender system problem.