Construction and Application of Predictive Model of College Students' Mental Health Based on Machine Learning

In recent years, faced with the increasingly frequent, complicated and hidden mental health problems of college students, colleges and universities at all levels have launched a variety of campus mental health education measures to realize timely discovery and early intervention of psychological problems. Among them, the commonly used methods of monitoring and early warning of college students' mental health at this stage still rely on psychological scale evaluation, psychological interview, etc., which have obvious shortcomings in the accuracy and timeliness of prediction results and are difficult to meet the practical application needs of colleges and universities. In this regard, this paper will focus on analyzing the application feasibility of machine learning technology in this field, and put forward a brand-new prediction model of college students' mental health, so as to improve the scientific and effective mental health education in colleges and universities. Practice has proved that the model is based on machine learning technology, and uses FP-Growth algorithm to mine the potential association rules between multi-source data such as students' behavior, learning performance and social relations, and realizes the monitoring and early warning of college students' mental health through the high accuracy and fast prediction ability of XGBoost algorithm, thus providing scientific basis and technical support for mental health education in colleges and universities.