College Student Lifestyle Query Classification Using Multi-Model Ensemble Learning with Polling Technique

Lifestyle-related issues among college students are mostly ignored by family and peers. If the queries related to lifestyle issues are not addressed in time, it aggravates serious issues, like depression, and leads to suicidal symptoms. This paper presents a multi-model classifier system with ensemble learning and polling technique to classify the student’s lifestyle queries. The primary dataset is generated with the lifestyle queries of college students aged between 17 and 23 years. The dataset is simulated with nine machine learning algorithms, and the ensemble learning model is build using the best four algorithms. This multi-model ensemble technique with a polling scheme has shown good performance in comparison with single-machine learning model classifier systems. The overall model has achieved an average precision score of 79%, recall score of 69% and F1-score of 72%.