BP Neural Network Feature Selection Based on Group Lasso Regularization

In this paper, the feature weight coefficient of neural network is proposed based on the results of neural network feature selection. Firstly, this paper introduces the neural network feature selection method based on Group Least absolute shrinkage and selection operator (Lasso) penalty, smooth approximation technique and the feature weight coefficient. Then, a two-stage algorithm for back propagation (BP) neural network feature selection based on Group Lasso penalty term is proposed. And then, the paper compares and analyzes the influence of smooth approximation technique on BP neural network feature selection. Finally, the reliability of the algorithm and the validity of the feature weight coefficient are verified by comparing the results of four datasets.