The Structure of Mental Elasticity Education for Children in Plight Using Deep Learning

The purpose is to solve the problem that the current research on the impact of the microstructure of mental elasticity and its constituent factors on the development of the mental elasticity of children is not comprehensive, and the traditional artificial analysis method of mental problems has strong subjectivity and low accuracy. First, the structural equation model is used to study the microstructure of poor children's mental elasticity, and to explore the structural relationship and functional path between the mental elasticity of children and the self-efficacy of their mental health, psychological anxiety, and attachment. Second, a prediction model of mental problems of children in plight based on the backpropagation neural network (BPNN) is constructed. Finally, middle schools in the representative areas of Northwest China are selected as the research unit. The relevant research data are collected by issuing questionnaires, and the data set is constructed to verify the performance of the model. The experimental results show that the average prediction errors of the BPNN model and the support vector regression (SVR) model are 1.87 and 5.4, respectively. The error of BPNN is 65.4% lower than that of SVR, so BPNN has a better performance. The prediction results of the test set show that the actual error and the relative error of the BPNN model are controlled in the range of 0.01, and the prediction accuracy is high. The structural equation model has a high fitting degree. The results of the questionnaire analysis show that attachment, self-efficacy, and psychological anxiety exert a significant direct impact on mental elasticity. This exploration aims to conduct a micro investigation on the relationship among the three core variables (attachment, self-efficacy, and mental health) in the resilience research of children in plight, and analyze their resilience, to provide a theoretical basis for the resilience intervention design of vulnerable groups.