Decision Tree-based Early Warning Model Of College Student Psychological Crisis

In recent years,the number of college students suffered from psychological problems is surging,which has been giving rise to some unfortunate results – suicide,the ‘major killer' for young people aged from 15 to 34 and has brought severe trauma both to families and society.Thus,it is of great significance for colleges to refine psychological crisis warning system and intervention analysis for students.Currently,the usual psychological crisis warning method is relatively single.SCL scale test,a conventional one widely used at the entrance stage for students,is less of efficiency.Therefore,it is vitally important for student managers to make effective use of new technologies to collect and interpret the students' personal attributes data,so as to get a clear psychological crisis warning in time.This paper proposes a decision tree based psychological crisis warning system for college students.After evaluation,the C4.5 classification algorithm is introduced to analyze the significant characteristic attributes that may cause psychological crisis,and different values of the attributes can tell whether the student may have psychological crisis or not.Throughout the selecting of the most significant feature attributes,constructive suggestions from expert on college student management are taken into consideration and binary logistic regression analysis are used,and based on the analysis six main attributes are selected: character,family members,household income,family relationships,absence reason and academic performance.Then,this paper looks in detail at the 14 rules for the positive early warning results which are used to describe the model.According to relevant research,it clearly indicates that the model has great credibility and reference.The test on different dataset shows that the accuracy of the prediction results can reach above 95%,precise rate is above 80%,recall rate reaches 60%,and the F value is stable at 70%,proving that the model is general and stable.In order to use the model in an effective way,centered on the early warning model,an application system is designed and implemented in this paper.It is a web-based B/S system,using the Spring MVC framework in the background,JSP in the front end and My SQL for the database.The system mainly implements basic functions such as collecting students' basic information,grade management,attendance record,and the corresponding data generated by this system will support the psychological crisis early warning system to offer helpful suggestions for grassroot student management work eventually.