摘要
为进一步提升心理测试的效果,提出一种基于单片机和SVM的高精度的心理测试仪。其中,首先采用STM32和D8233芯片实现生理信号的采集,然后进行ECG和HRV特征提取,并将提取的特征输入到萤火虫算法优化的支持向量机心理健康评估模型中。实验结果表明,通过单片机的测试仪对生理信号的采集,可提取出ECG和HRV生理特征;通过不同特征下的心理健康评估得出,使用14维HRV特征能取得较好的分类效果,在SVM和FA-SVM算法上分类准确率分别达92%和97%;与其他算法相比,提出的FA-SVM心理健康评估在Augsburg情感生理数据集上的分类准确率达95.88%,F1-Score值达0.9238。以上结果表明,所构建的基于单片机的心理测试仪进行心理测试时能够取得良好的分类效果,具有一定的实际工程价值。
In order to further improve the effect of psychological test,a high precision psychological tester based on SCM and SVM is proposed.Specifically,STM 32 and D8233 chips were used to collect physiological signals,followed by ECG and HRV feature extraction,and the extracted features were input into the SVM mental health assessment model optimized by Firefly algorithm.The experimental results show that the physiological features of ECG and HRV can be extracted from the physiological signals by the mental health assessment under different features,the 14-dimensional HRV features can achieve the classification accuracy of 92% on SVM and FA-SVM and 97% respectively.Compared with other algorithms,the FA-SVM mental health assessment on the Augsburg emotion physiology data set is 95.88%and the F1-Score value is 0.9238.The above results show that the CM-based psychological tester can achieve good classification effect and have certain practical engineering value.
作者
杨伟樱
兰天
刘月
李晔
吕阿璐
康北莉
YANG Weiying;LAN Tian;LIU Yue;LI Ye;LV Alu;KANG Beili(Shaanxi Polytechnic Institute,Xianyang Shaanxi 712000,China;Xianyang Psychological Association,Xianyang Shaanxi 712000,China)
出处
《自动化与仪器仪表》
2023年第5期235-239,共5页
Automation & Instrumentation
基金
陕西省教育厅2022年度一般专项科研计划项目(自然科学项目)《基于大数据的高职学生心理档案整合与动态心理健康监测系统研究》(22JK0264)。
