An In-class Teaching Comprehensive Evaluation Model Based on Statistical Modelling and Ensemble Learning

Abstract

Education is the foundation of our cities' and society's development. The realization of Smart City is inseparable from the construction of Smart Education. The development of science and technology has promoted the popularity and booming of information technology in the education field. With the rise of Smart Education, intelligent learning and evaluation have provided new ideas for in-class teaching evaluation. The model proposed in this article is based on existing multi-dimensional and multi-modal data set from in-class audio and video recognition, as well as movement perception and interaction analysis. We firstly designed a statistical model and an ensemble learning model for in-class teaching evaluation, which are based on Analytic Hierarchy Process - Entropy Weight Method and AdaBoost algorithm respectively. Then, we designed experiments to assess the performance of the proposed statistical model and ensemble learning model. Finally, we compared and selected better models through experiments in different evaluation indicators and combined them into our In-class Teaching Comprehensive Evaluation Model with outstanding performance.