A method proposed for estimating depressed feeling tendencies of social media users utilizing their data

Abstract

The enormous progress of social networks and the large amount of data generated by them, has led many studies the possibility to identify the hidden knowledge. Depression is more than just feeling unhappy or fed for a few days and affects people in different ways and can cause a variety of symptoms. In this paper, we aim to analysis on research for feelings of depression for user’s activities in Twitter, which is a popular microblogging site, for estimating his/her depressive tendency. Then we investigate Weka as a tool of machine learning classification to extract useful information for classification of Twitter Data collected from Twitter based Twitter API. Therefore, we perform our experiments to estimate participants depressive tendencies using an algorithm for computing the semantic similarity between tweets in training set and data set based on WordNet as an external semantic resource. Experimental results show that Twitter could be used to analyze online depression feeling. In addition, this study demonstrated that we can extract sentiments of Twitter users from social networks.