LAUSR

The intelligent and automatic evaluation for teaching quality has been a more and more general demand in the digital society. The most intuitive solution is to make an assessment based on examination results, which cannot comprehensively reflect the course situation. To deal with such a problem, this paper designs a smart knowledge discovery system for teaching quality evaluation using a genetic algorithm-based BP neural network. Three aspects of factors are selected as the features: teaching conditions, teaching process, and teaching effect. The BP neural network is formulated to learn a nonlinear mapping from initial features to evaluation results. The learned parameters are then re-optimized by introducing the genetic algorithm. The proposed evaluation framework is also assessed concerning running performance on a real-world course-teaching dataset. The obtained results reflect that the proposal can realize the intelligent evaluation of the teaching effect and the evaluation effect is close to artificial experience-based evaluation results.