Construction of Daily Screening and Early Warning System for Students with Psychological Abnormalities Based on Big Data and Deep Learning Models 
Xu Hao

Sichuan Institute of Industrial Technology，Sichuan ProvinceDeyang City，618500；

Abstract：Mental health education in colleges and universities in the new era is confronted with practical dilemmas such as the frequent occurrence of students’ psychological problems, strong concealment of psychological risks, and the lag of traditional screening methods. Relying on subjective scale evaluation, traditional psychological screening is plagued by many shortcomings including single data dimension, strong subjectivity, poor timeliness and delayed early warning, which make it difficult to adapt to the complex and changeable needs of college students’ psychological management. To solve the above problems, this study relies on big data and deep learning technologies, takes the Temporal Convolutional Network (TCN) as the core model, integrates multi-source campus objective behavioral data, and constructs a daily screening and early warning system for college students with psychological abnormalities.Firstly, the research sorts out the policy background and practical pain points of current college mental health work, analyzes the limitations of the traditional psychological screening model, and expounds the value of system construction from both theoretical and practical dimensions. Secondly, it clarifies the advantages of the core technical model, optimizes the whole process of multi-source data collection, processing and analysis, and establishes a scientific and intelligent psychological screening and early warning framework. Finally, it improves the application mechanism and implementation path of the system. The research results show that the intelligent early warning system based on big data and deep learning can effectively make up for the deficiencies of traditional screening methods, accurately capture implicit psychological abnormal signals of students, and realize the early detection, early warning and early intervention of psychological risks. It provides technical support and practical reference for the scientific, intelligent and normalized development of college mental health education.

Keywords: Big Data; Deep Learning; TCN Model; College Students’ Psychological Abnormality; Screening and Early Warning System

References

[1]Shukla A, Anand S. A review of temporal convolutional networks for sequence modeling[J]. International Journal of Business Intelligence and Data Mining, 2019, 14(1-2): 193-202.

[2]Lea C, Flynn M D, Vidal R, et al. Temporal convolutional networks for action segmentation and detection[C]//Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition. 2017: 156-165.

[3]Liu S, Young S D. A survey of social media data analysis for physical activity surveillance[J]. Journal of Forensic and Legal Medicine, 2018, 57: 33-36.

[4]Xu L D, He W, Li S. Internet of things in industries: A survey[J]. IEEE Transactions on Industrial Informatics, 2014, 10(4): 2233-2243.

[5]Hunt J, Eisenberg D. Mental health problems and help-seeking behavior among college students[J]. Journal of Adolescent Health, 2010, 46(1): 3-10.

[6]Downey L A, Rapport P J. Using the psychological well-being model to predict depression among Australian university students[J]. Australian Journal of Psychology, 2012, 64(1): 38-47.

[7]Reavley N J, Jorm A F. Prevention and early intervention to improve mental health in higher education: a review[J]. Early Intervention in Psychiatry, 2010, 4(1): 3-10.

[8]Zhu Yan, Lin Shuang. Construction of college students’ psychological early warning system from the perspective of big data[J]. Coal Higher Education, 2020, 38(01): 60-67.

[9]Hunt J, Eisenberg D. Mental health problems and help-seeking behavior among college students[J]. Journal of Adolescent Health, 2010, 46(1): 3-10.

[10]Zhao Yifan, Li Ling, Li Pengsen, et al. Hotspots and frontiers of psychological crisis intervention research in China: Based on CiteSpace visual analysis[J]. Ergonomics, 2022, 28(06): 45-50.