Construction of teaching game evaluation model based on ISSA-BPNN

Abstract

Teaching games are an effective teaching organization activity. In response to the evaluation and prediction problem of teaching games, a teaching game evaluation model based on improved sparrow search algorithm and back propagation neural network was studied and constructed. Firstly, a situational teaching game was designed and an evaluation index system was constructed. Then, a teaching game evaluation prediction model based on the improved method was established. Finally, the expert consultation method is adopted to collect opinions from experts in the field of education and construct an evaluation index system for teaching games. And based on the evaluation index system of teaching games, evaluate students’ mathematical thinking ability before and after experiencing teaching games to verify the application effect of teaching games. The scenario based teaching game designed in this study has a certain effect on improving students’ mathematical thinking ability. Students’ mathematical thinking has significantly improved (P<0.05), and the teaching effect is the same for students of different genders (P>0.1). The improved sparrow search algorithm has a faster convergence rate than other algorithms, and tends to be stable when iteration is about 100 when solving the single peak benchmark function. When solving the multimodal benchmark test function, it tends to stabilize when iteration is around 20. The teaching game evaluation prediction price model based on the improved method shows a trend of first increasing and then decreasing with hidden units increasing. When the hidden unit is 16, the area index under model curve is the highest, around 0.962, and its prediction accuracy is relatively high. In summary, the model constructed in this study is applicating good in teaching game evaluation prediction, and can promote education industry developing.