基于多维度特征的大学生心理健康状态评估  

Evaluation of College Students Mental Health Based on Multi-Dimensional Characteristics

作　　者：蔡婉君[1] 姚俊[2] CAI Wanjun;YAO Jun(Mental health consulting department,ZhangZhou Institute of Technology Intelligent manufacturing College,ZhangZhou Institute of Technology,ZhangZhou,FuJian 363000,China)

机构地区：[1]漳州职业技术学院心理健康咨询科 [2]漳州职业技术学院智能制造学院,福建漳州363000

出　　处：《九江学院学报（自然科学版）》2022年第4期7-12,共6页Journal of Jiujiang University：Natural Science Edition

基　　金：福建省中青年教师教育科研项目(编号JAS171085)的研究成果之一。

摘　　要：以大学生行为特征为基础,进行心理健康状态评估时,主要依据单一维度的特征,使得评估结果均方误差较大。因此,文章引入多维度特征理念,设计一种新的大学生心理健康状态评估方法。引入信息熵概念对心理状态感知数据进行分析,完成数据特征维度的构造与选择。结合常规心理状态自评标准,建立心理健康评估指标体系,并为每项评估指标进行赋权。通过局部线性嵌入算法对多维度特征数据进行降维处理,再结合支持向量机算法,构建基于多维度特征的评价模型。最后,运用模糊综合算法和最大隶属原则,得到大学生心理健康评估结果。实验结果表明:所提方法评估结果的均方误差为0.02,与基于决策树、基于前馈神经网络的评估方法相比,均方误差分别降低了11%、17%。Based on the behavioral characteristics of college students,the assessment of mental health status was mainly based on the characteristics of a single dimension,which made the mean square error of the assessment results larger.Therefore,the concept of multi-dimensional characteristics was introduced to design a new method to evaluate the mental health status of college students.The concept of information entropy was introduced to analyze the psychological state perception data,and the construction and selection of data feature dimensions are completed.Combined with the conventional self-assessment criteria of mental state,the mental health assessment index system was established,and each assessment index was weighted.The local linear embedding algorithm was used to reduce the dimension of multi-dimensional feature data,and then combined with the support vector machine algorithm,an evaluation model based on multi-dimensional features is constructed.Finally,the fuzzy comprehensive algorithm and the maximum membership principle were used to get the results of College Students'mental health assessment.The experimental results showed that the mean square error of the evaluation results of the proposed method was 0.02.Compared with the evaluation method based on decision tree and feedforward neural network,the mean square error was reduced by 11%and 17%.

关 键 词：多维度特征 大学生 心理健康 状态评估 局部线性嵌入 支持向量机 

分 类 号：G641[文化科学—高等教育学]