基于改进ANN的心理状态预警建模与仿真

Modeling and simulation of psychological state early warning based on improved ANN

针对现有算法对心理健康状态评估准确度低,难以有效替代传统人工处理方式,导致心理危机预警不及时等问题,文中设计了一种基于改进ANN的心理危机预警算法模型。该模型在全面收集可能影响学生心理状态信息的基础上,通过筛选与特征提取分析得到了ES-ANN网络所需的数据集合,再引入BP算法与ReLU激活函数获得相关预警模型。该模型可在被测者心理存在异常问题时及时发出预警信息。实验结果表明,该算法对心理危机预警的准确率及查准率能够分别达到92.1%和81.6%,且均优于其他对比算法。

Aiming at the problems that the existing algorithms have low accuracy in the evaluation of mental health,are difficult to effectively replace the traditional manual processing methods,and lead to the untimely early warning of psychological crisis,this paper designs a psychological crisis early warning algorithm model based on improved ANN.Based on the comprehensive collection of information that may affect students'psychological state,the model obtains the data set required by ES-ANN network through screening and feature extraction analysis,and then introduces BP algorithm and ReLU activation function to obtain the relevant early warning model.The model can send out early warning information in time when the subjects have abnormal psychological problems.The experimental results show that the accuracy and precision of this algorithm for psychological crisis early warning can reach 92.1%and 81.6%respectively,and are better than other comparative algorithms.