Research on Classification Model of BP Neural Network Based on DL Algorithm

To accurately and quickly find out the degree of depression tendency for teenagers, so people can make timely intervention work to reduce the incidence of psychological diseases of them. In this paper, we present a BP neural network classification model based on DL algorithm. Comparing with the steepest descent algorithm and conjugate gradient method of BP neural network, the convergence speed of DL algorithm is higher than them in the inversion coefficient experiment of the equation. And its accuracy is similar to that of conjugate gradient method, but much better than that of steepest descent algorithm. In the experiment of Chinese text sentiment analysis, the precision, recall and F1 values of DL-BP are better than those of SD-BP and CG-BP. Based on DL-BP model, the early warning platform of teenager depression tendency is designed in this paper. It is a design with good stability, reliability and extensibility.