Identification and Modeling of College Students’ Loneliness from the Perspective of Big Data a Machine Learning-Based Approach

Abstract

With the advent of the era of big data, the information society inevitably intersects and integrates with everyone's life. Compared with the traditional selfstatement scale for psychological measurement, big data network information has advantages such as excellent ecological validity. In this paper, we use the consumption data and access control data of college students in a university as the data source and the employment situation as the target variable to identify and predict the loneliness of college students. In the context of the current COVID-19 epidemic, which is generally in quarantine, this paper provides a realistic basis for this study. By extracting the features from the data, we address the limitations of the machine learning modeling approach for the autonomous identification and prediction of loneliness symptoms and propose further development prospects. This paper provides a practical basis for this research. By extracting features from the data information, we propose the limitations of the machine learning modeling approach for the autonomous identification and prediction of symptoms of loneliness, and propose the future development of this approach.