Abstract :
A social recommendation system has attracted a lot of attention recently in research communities. Traditional social recommendation algorithms are often based on batch machine learning methods that suffer from several critical limitations, for example, a cost of training is extremely expensive in each time new user ratings arrive, unable to capture the change of the users preferences over time. The proposed system is a new online social movie recommendation framework from the view point of online graph regularized user preference learning OGRPL , which incorporates the user item collaborative relationship as well as content characteristics of elements in a unified process of learning preferences. This method overcomes the problem of excessive data adjustment by eliminating the low range constraint. However, it does not handle uncertain ratings that contain irregularities. The propagation of the Monte Carlo uncertainty can be used to handle uncertain data in the preferences modeling. The level of uncertainty is measured. Also use SVM Support Vector Machine to predict user preferences using the model learning method. SVM is a classification method which is used to classify movies. Then based on classification recommendation is performed. Sreelekshmi. B | Vidya. N "Survey on Kernel Optimization based Enhanced Preference Learning for Online Movie Recommendation" Published in International Journal of Trend in Scientific Research and Development (ijtsrd), ISSN: 2456-6470, Volume-2 | Issue-3 , April 2018, URL: https://www.ijtsrd.com/papers/ijtsrd11305.pdf Paper URL: http://www.ijtsrd.com/computer-science/data-miining/11305/survey-on-kernel-optimization-based-enhanced-preference-learning-for-online-movie-recommendation/sreelekshmi-b
Keyword :
Online social recommendation, user preference learning, low rank matrix, uncertainty, over fitting problem