FRIEND MATCHING USING PROBABILISTIC TOPIC MODEL release_tx4fxd3arbbq3j62utab7whtri

Abstract

Recommender Systems provide suggestions for users to guide in various decision-making processes.The recommender systems can be defined by the purpose of recommendation, mechanism and data gathering. Recommendation system for social networks are different since the item recommended are rational human beings. The paper focuses on designing a friend matching system by analyzing user lifestyles as common criteria. Large amount of data collected from various users create high dimensional data. In order to resolve this, probabilistic topic modeling is used. Content based machine learning approaches are used to find out suspicious users in the recommendation system. The results are evaluated based on the datasets created from the real world users.
In text/plain format

Archived Files and Locations