Prediction model and case analysis of college students' psychological depression based on multi-source online comment mining

Abstract

Psychological depression is a normal emotional experience of human beings. Everyone will experience different levels of depression in life. Under the dual influence of the current socio-economic environment and the small environment of students' quality, the depressive tendency of college students cannot be ignored. In order to mine and improve the level of College Students' psychological depression, this paper proposes a prediction model of College Students' PD based on multi-source online comment mining. The data mining method is used to analyze the content and emotion of microblog comments of users with depressive tendencies. Then, pattern extraction and matching are used to find low-frequency feature words. The example analysis shows that when the comment length is set to 10 and the news length is set to 47, the classification accuracy of the test set is the highest, reaching 96.454%, higher than the original 94.898%. Learning pressure, economic pressure, employment pressure, coping style and social support are closely related to depression and anxiety. Therefore, when modeling depression and anxiety, they were selected as predictive properties. The PD prediction model of college students based on multi-source online comment mining has achieved good results in the polarity classification of online comments.