Cold Start Recommender System Using Tensor Based Methods

Abstract

In the era of internet, we are facing problem of data overload. Due to this every internet user is facing problems while doing any internet activity such as writing posts on blog, updating Facebook, while using chat messages/application and while getting email notifications and spam messages. Hence all these activities become unmanageable which cause less efficiency and low productivity in any system. One solution to this data overload problem is personalized recommendations, which helps online users, manage with big volumes of varied data on the Internet, many intellectual software has appeared which enable users to locate appropriate items that meet up their desires. Many recommendation systems are available to solve this purpose such as YouTube.com, twitter.com, Spotify and many more. Most of the systems are considering only two entities user and item while making prediction. Recommendation accuracy can be increased if we consider extra information. Since last few years, the multidimensional attributes which clarify user performance and likings have been getting growing interest. In this direction many research works are going on which works for two-dimensional data, but these methods do not give good result in case of multidimensional data. In parallel with this we have come up with next group of WWW, known later as Web 2.0. it has taken web to a different height. With web 2.0, any user can communicate and interact with other users. Now a days Social Tagging System (STS) has attracted many researchers. In STS all the users are allowed to interpret, explain, or gather information by giving input as each user s specific likings and desires. User can choose tags freely based on description provided by frequent users. Personal views given by the users helps in predicting user tagging behaviour. Social Network highly influences the opinion of users. newlineRecommendation systems consider that STS have the two major issues like cold start and data sparsity. Data Sparsity is arising when only few users are interested in ra