Résumé
Combining the psychological "big data" of college students with the data of the psychological expert database, using data mining algorithm to extract characteristic attributes, deriving the warning results from a stable algorithm model, the psychological early warning platform constructed by this method can not only monitor students' psychological dynamics in real time, but also predict their psychological behavior trend, which makes up for the lag and low accuracy of traditional early warning methods. The psychological early warning platform provides data support and the psychological early warning system provides operational guarantee. The psychological early warning mechanism formed by the combination of the two can provide scientific basis for educators and finally achieve the goal of using big data technology to improve the mental health of college students.