Social Behavior Prediction using Ensemble Learning Algorithms

In this modern era, social media users are spreading daily. Most social media users spend much time on these platforms daily. While using social networks, people can connect within a few seconds and attain information from any part of the world. Nevertheless, using social networks also have some adverse effects. Recently, researchers have shown that social networks negatively affect our behavior. In this work, a method was presented to analyze social behavior. Our analysis was started with data collection in the form of questionnaires from different individuals. The dataset was preprocessed by handling missing values and applying SMOTE technique. After that, Genetic search for the feature selection technique, ranking features with Information Gain, and Chi-square test with the p-value were used. Finally, three machine learning algorithms were applied for the classification algorithm with Rotation Forest, Dagging, and Decorate. In our analysis, Decorate ensemble classifier achieved the satisfactory performance.