An application of neural network algorithm model based on improved multi-expression programming in English language education practice

Abstract

In the field of English education, neural network algorithm can effectively predict and evaluate teaching, and significantly improve the quality of education and teaching. Therefore, a neural network English teaching evaluation prediction model based on multi expression programming is proposed. Through the research on neural network and genetic algorithm (GA), it is found that flexible neural tree cannot optimise parameters and results at the same time. Therefore, a neural network algorithm model (MEP) based on multi expression programming is proposed to solve the problem, and the MEP-NN English teaching evaluation model is constructed by optimising the model parameters with evolutionary algorithm. The model is applied to the English teaching process to achieve the evaluation of English teaching quality. The results show that in the mean square error performance test of multiple algorithms, achieving convergence after 500 iterations, with an MSE value of 0.02 and the best error performance; in the English class comprehensive quality prediction, the proposed MEP-NN algorithm has the best prediction accuracy, with a prediction mean of 86.56 points, closest to the actual value of 86 score, with a prediction accuracy of 94.56%. This shows that the proposed MEP-NN algorithm has excellent performance.