Analysis model of college students' mental health based on online community topic mining and emotion analysis in novel coronavirus epidemic situation

Under the epidemic situation of COVID-19, university students have different levels of anxiety, depression, and other psychological problems, and these differing levels present different challenges. Therefore, universities and relevant departments should carry out accurate psychological health education for university students. Through research, this paper found that students' psychological problems during the COVID-19 epidemic were mainly reflected in four aspects: depression, interpersonal relationship, sleep and eating disorders, and compulsive behavior. Through the discussion of family of origin, self-awareness and motivation attribution, and social pressure, this paper analyzed the causes of psychological problems. The information resources of the network are usually unstructured data, and the text information, as the most typical unstructured data, occupies a large proportion. Moreover, this text information often contains users' emotional response to major events. In this paper, a data preprocessing system is designed, and three data preprocessing rules are defined: expression data conversion rules, data deduplication rules and invalid data cleaning rules. The characteristics of online community text data are analyzed, and the text feature extraction method is selected according to its characteristics. The results of this study show that the proportion of university students with psychological problems is about 23%, which is slightly higher than the research results during the non-epidemic period. This paper suggests that college students should master methods of self-regulation, improve their levels of physical exercise, improve their physical fitness, and establish and improve their defense mechanisms to alleviate psychological conflicts and pressures.