Trending Topics Detection using Machine Learning Approach

In this age the social media is one of such kind of application where easily data is obtainable by the users. According to these the test of user's can also be identifiable. Due to this the application of text mining is also increased in social media text mining and analysis. In this presented work social media text analysis based technique is introduced. That technique helps to identify the most trending topic from the micro-blog data. In this context three things are necessary first social media data, creation of topic model and evaluation of text data based on developed data model. Therefore for an effective social media data twitter data is considered. In further the preprocessing techniques are applied for quality improvement of raw data. In next the regular size of data is extracted as features of text twits using the term frequency and sentence formation probability. Finally the clusters of data are computed using FCM (fuzzy c means) clustering and the clustered data is used with the bay's classifier for assuring the topic name. The implementation of this topic model is performed on JAVA technology. After that on the basis of experimental evaluation the performance of topic model is computed. Additionally a comparative analysis with traditional topic model is performed. The computed outcomes show the proposed model is an efficient and accurate technique of data analysis.