Application of Hybrid GA-SOFM Neural Network in Quality Evaluation of English Teaching

Abstract

SOFM neural network algorithm adopts an unsupervised clustering algorithm, which can map the cluster center generated after calculation to a surface or plane, which makes the topology of the network have high stability. The GA algorithm completes the operation process through three operators: selection, crossover, and mutation. It has good global optimization and robustness. In this paper, the SOFM algorithm is improved by a GA algorithm and a hybrid GA- SOFM neural network algorithm is established. The algorithm is applied to the quality evaluation system. According to the results of the MATLAB simulation experiment, the evaluation accuracy and absolute error are determined and compared with the previous optimal GA-RBF hybrid algorithm model. The results show that the average evaluation accuracy of the proposed algorithm model evaluation is 89.43%, and its absolute error is 0.017. It shows that the quality evaluation model based on a hybrid GA-SOFM neural network can effectively and accurately evaluate quality.