UNDERSTANGING THE USER PREFERENCES USING SOCIAL NETWORK MINING

Abstract

Recommender systems have become an important part of the web sites; the vast number of them is applied to e-commerce. They help people to make decision, what items to buy, which news to read, or which movies to watch. Recommender systems are particularly useful in environments with information overload since they cope with selection of a small subset of items that appear to fit the user’s preferences. The global network provides a vast amount of diverse data useful for social network analysis, e.g., for the estimation of the user social position or finding significant individuals or objects. Internet-based social networks can be either directly maintained by dedicated web systems like Twitter, Facebook , LinkedIn or extracted from data about user activities in the communication networks like e-mails, chats, blogs, homepages connected by hyperlinks , etc. Some researchers identify the communities within the Web using link topology, while others analyze the e-mails to discover the social network. Based on semantic web analysis and using soft computing techniques and data mining tools the relevant information is obtained from the social network and by applying the different techniques and approaches of data mining and soft computing, data can be clustered to be an input to the Recommender system. The main focus of this thesis is extracting the relevant information from the social network site like Twitter using SNM and designing an appropriate RS for understanding the user preferences.