摘要
大众心理健康问题日益受到广泛关注。为提高心理医疗资源的使用效率,采用互联网手段采集与心理咨询相关的数据,使用一种融合脚本筛选、数据整理以及利用Easydata进行主动学习式数据标注的新方法构建数据集,结合卷积神经网络和BERT预训练模型等技术,在textvec-base-chinese模型的基础上,提出了PycModel模型,以实现更加高效的心理咨询问题分类。实验结果显示,PycModel在心理咨询问题分类的准确率明显优于其他参照模型,该模型能够有效提高心理咨询的效率和效果,可以为心理健康服务提供有力的支持。
Issues concerning public mental health are receiving widespread attention.To improve the efficiency of psychological medical resource utilization,a new method that integrates script filtering,data organization,and active learning-based data annotation using Easydata is adopted to collect data related to psychological counseling via the internet.Combining techniques such as convolutional neural networks and BERT pre-trained models,and building upon the textvec-base-chinese model,the PycModel is proposed to achieve more efficient classification of psychological counseling issues.Experimental results show that the PycModel significantly outperforms other reference models in terms of accuracy in classifying psychological counseling issues.This model can effectively enhance the efficiency and effectiveness of psychological counseling,providing strong support for mental health services.
作者
易云恒
张超群
武家辉
汤卫东
Yi Yunheng;Zhang Chaoqun;Wu Jiahui;Tang Weidong(Department of Information Engineering of Hope College,Southwest Jiaotong University,Chengdu,Sichuan,China 610400;School of Artificial Intelligence,Guangxi Minzu University,Nanning,China 530000)
出处
《深圳信息职业技术学院学报》
2024年第4期56-64,共9页
Journal of Shenzhen Institute of Information Technology
基金
国家自然科学基金项目(项目编号:62062011)
广西壮族自治区大学生创新创业项目(项目编号:S20211060821S)。
关键词
文本分类
心理咨询问题分类:深度学习
text classification
classification of psychological counseling problems
deep learning
