Identify User Behavior Based on The Type of Tweet on Twitter Platform Using Gaussian Mixture Model Clustering

Social media has now become a place for social interaction to exchange information about business, politic, and many other. Twitter is one of the platforms that provides services their users share opinions on certain topics. The topic will be discussed in this study politic by collecting tweet data student demonstration movement SemuaBisaKena campaign. By using word weighting method TF-IDF Vectorizer Gaussian Mixture Model Clustering, it possible identify whether user behavior positive (support) or negative (blasphemy). To achieve final result, there are several stages must passed. Such as preprocessing, feature extraction Vectorizer, Clustering algorithm visualization. results 1 cluster identified 2 clusters behavior.