基于大数据的大学生心理精神状况分析与预测  

Analysis and Prediction of College Students’Mental State Based on Big Data Analysis

作　　者：李清 邓国英[2] 苏强[3] LI Qing;Deng Guoying;Su Qiang(School of Economics and Management,Shanghai University of Political Science and Law,Shanghai 201701,China;Trauma Center of Shanghai First People’s Hospital,Shanghai 201620,China;School of Economics and Management,Tongji University,Shanghai 200092,China)

机构地区：[1]上海政法学院经济管理学院,上海201701 [2]上海市第一人民医院创伤中心,上海201620 [3]同济大学经济与管理学院,上海200092

出　　处：《运筹与管理》2024年第12期210-216,共7页Operations Research and Management Science

基　　金：国家自然科学基金面上项目(71972146);上海市哲学社会科学规划教育学青年项目(B2024001)。

摘　　要：基于SCL-90心理健康测试量表,本文选取抑郁、焦虑、强迫、偏执和人际关系敏感5个分量表,分析大学生的心理精神状况。面向全国大学生发放问卷,收回有效问卷8664份,采用卡方检验分析对大学生心理精神状况有显著影响的指标。根据分量表得分将大学生心理分为较弱、一般和较强三类,采用决策树、随机森林、多层感知机和支持向量机4种机器学习方法对大学生心理精神状况进行预测;在支持向量机中,采用高斯核函数构建模型,并使用粒子群算法优化核函数参数。5个分量表下,核函数参数优化后支持向量机的预测准确率最高。With the rapid development of society,the increase in cost of living and pressure of study and work,the psychological problems of college students become increasingly prominent.At present,teenagers all over the world have psychological problems to a certain degree.College students’psychological state affects all aspects of study and life,and the impact is increasingly intensified.Mental problems can also lead to physical problems such as physical pain and visual fatigue.Therefore,it is the primary task of schools and families to accurately identify college students with psychological problems and improve their psychological quality and anti-pressure ability.Applying big data technology to the field of students’psychology and spirit can collect and analyze all aspects of students’information,dynamically track students’behavior,and capture abnormal information and behavior in real time,so as to intervene in time.Based on SCL-90 scale,a questionnaire is designed to investigate the psychological states of college students.In this paper,five subscales-depression,anxiety,compulsion,paranoia and interpersonal sensitivity,are selected to analyze the psychological states of college students.According to the score of the scale,the psychological states are divided into three grades:weak,general and strong.The frequency of visual fatigue and physical pain under each grade is analyzed,and the relationship among visual fatigue,physical pain and the five mental states is analyzed by the Chi-square test.Four prediction methods,namely decision tree,random forest,multi-layer perceptron and support vector machine,are used.Particle swarm optimization is used to optimize kernel function parameters of support vector machine to improve the prediction accuracy.(1)14%-32%of the students have a general degree of depression,anxiety,etc.,more than 70%of whom suffer from visual fatigue.3%-10%have strong symptoms of depression,compulsion and other symptoms,80%of whom suffer from visual fatigue.Psychological conditions such as depressi

关 键 词：大数据 大学生心理精神状况 机器学习 支持向量机 粒子群算法 

分 类 号：TP181[自动化与计算机技术—控制理论与控制工程]