Cut-based classification for user behavioral analysis on social websites

Everyday millions of users can share or exchange their opinion through messages on social web sites. In various domains behavior analysis is critical for decision making. The behavioral data on social website can provide an economical and effective way to expose public opinion timely. The public behavior in messages can be used to obtain user feedback towards different company products; it can be utilized for marketing of different products or to track the popularity of different things. So there must be some methodologies to analyze user behavioral variations on social web sites and extract possible reasons behind such variations. This paper focuses on the analysis of user behavior on social networking sites. The analysis is carried out over various messages or text exchanged on the social network. Such analysis is helpful for taking certain decisions related to the trend prediction. The evaluation of the proposed methodology for behavior analysis is carried out in this paper on the basis of identified sentiments from the exchanged messages. A cut-based classification approach is proposed for analyzing the user's behavior in this paper.