Stacking Ensemble Learning-based Gender Identification for User Profiling in Smart Education

To improve company profit, user profiling has become an important research topic in understanding customers and improving products or services, in which identifying users' gender is a fundamental problem. However, little research work was carried out on user profiling for improving learning in the domain of Smart Education. By analyzing Chinese microblog data, this paper proposes a stacking ensemble learning-based gender identification method for user profiling. It extracts gender features according to user published microblog messages and uses the base learner to extract stacking features, which are used as input to the meta-learner to obtain gender identification results. Its effectiveness is verified by comparison experiments with existing algorithms such as Logistic Regression and Random Forest for social media user gender identification.