摘要
大学生通常需要面对更繁重和专业的课程学习,同时还需要应付考试和就业压力等,使其容易出现心理健康问题。为了及时监测干预大学生的心理健康状态,研究设计了一种基于智能可穿戴设备的心理健康数据自动化采集方案。该方案采用智能可穿戴设备对学生心理健康数据进行采集,再利用主成分分析法对心理健康数据特征提取与降维,同时结合逻辑回归与Boosting集成学习方法构建高精度分类器。实验结果表明,研究所设计方案分类正确率达87.50%,准确率、精确率、召回率和k交叉验证预测准确率分别为0.97、0.95、0.94,显示出优越性能。并且该方案在数据采集与传输中表现出高稳定性和准确性,丢包率约为3%,误差率控制在0.2%以下。研究为大学生心理健康监测提供了一种可靠的技术手段,具有重要的实际应用价值。
College students usually need to face more demanding and specialized coursework,as well as cope with exam and employment pressures,making them prone to mental health problems.In order to monitor and intervene in the mental health status of college students in a timely manner,a research has designed an automated collection scheme for mental health data based on intelligent wearable devices.This scheme adopts intelligent wearable devices to collect students'mental health data,and then uses principal component analysis to extract and reduce the dimensionality of mental health data features.At the same time,it combines logistic regression and Boosting ensemble learning methods to construct a high-precision classifier.The experimental results show that the classification accuracy of the designed scheme by the research institute reaches 87.50%,and the accuracy,precision,recall,and k-cross validation prediction accuracy are 0.97,0.95,and 0.94,respectively,demonstrating superior performance.And the scheme exhibits high stability and accuracy in data collection and transmission,with a packet loss rate of about 3%and an error rate controlled below 0.2%.The research provides a reliable technical means for monitoring the mental health of college students,which has important practical application value.
作者
田嘉敏
苟仲荣
TIAN Jiamin;GOU Zhongrong(Xi’an Siyuan University,Xi’an 710000,China;Xi’an Vocational and Technical College,Xi’an 710000,China)
出处
《自动化与仪器仪表》
2025年第10期96-99,共4页
Automation & Instrumentation
基金
2024陕西省教育厅专项科研计划项目《AIGC助力陕西省基础教育阶段教师教学技能提升研究》(24JK0184)
校级项目《基于知识图谱的一流课程〈幼儿园班级管理〉》。
关键词
智能可穿戴设备
学生心理健康
数据采集
数据分类
特征提取
intelligent wearable devices
student mental health
data collection
data classification
feature extraction
