Graph Neural Network on Psychological Prediction of College Students Special Education.

With the closer life rhythm of college students and closer contact with social life, these aspects may increase the pressure of college students' life. Some students have not experienced professional psychological counseling, which will cause extreme psychological stress. When faced with a variety of everyday life demands such as study, love and friendship, job, and future career design, college students should have an optimistic, positive, and courageous mentality. Most of the problems of mental health are the result of college students' excessive stress, trivial family life events and other social factors. In this paper, a spatiotemporal graph neural network college student psychological prediction model for information augmentation transmission is proposed and an information augmentation transmission mechanism is designed to capture the complex dependencies of data. The adaptive graph learning structure can dynamically learn the dependencies between nodes and the enhanced information transmission mechanism can increase the area of information transmission to capture richer spatial information and select the effective information through the attention mechanism, and eliminate the invalid information; and design a data linear and nonlinear fusion mechanism, the mechanism includes linear and nonlinear parts, the two parts capture the linear and nonlinear characteristics of the data, and then Fusion.