In recent times, we have seen an explosion in the growth and popularity of social networking, which resulted in its problematic usage. There is an increase in number of social network addiction. Symptoms of these addictions are observed passively today, resulting in affecting the users adversely. In this paper, we propose that mining online social behavior provides an opportunity to monitor the addictive usage of the user. It is challenging to detect behavior because the mental status cannot bedirectly observed from social activity logs. Here, we propose a deep learning framework that exploits features extracted from social network data to accurately self -analyze the behavior in online social network users. We perform a feature analysis, and also machine learning on large datasets and analyze the characteristics of the user. Sentiment Analysis is used to identify and study affective states and subjective information. The retrieved result is displayed to the user in the form of statistical graphs.