基于深度学习的在线健康社区抑郁症用户画像研究

Research on the Depression User Profile of Online Health Community Based on Deep Learning

人工智能与深度学习技术为精准识别在线健康社区抑郁症患者奠定了基础.首先构建了基于TCNN-GRU深度学习的抑郁情感分类模型,进行在线健康社区实验数据集进行抑郁情感分类标注后,通过TCNN-GRU模型判别用户的抑郁症倾向;在此基础上,进一步提出抑郁指数的概念,通过对抑郁指数和患者抑郁程度两者关系的深度挖掘,由此建立基于深度学习的在线健康社区抑郁症用户画像模型.实验结果表明,与传统的卷积神经网络模型、循环神经网络模型以及混合模型相比,TCNN-GRU模型在抑郁情感分类上能获得了更优的结果,基于深度学习的在线健康社区抑郁症用户画像模型也能够从文本分析的角度准确识别用户的抑郁情感和抑郁状态.

Artificial intelligence and deep learning technology laid a foundation for the identification of depression patients in online healthy community.In this paper,we proposed a classification model of depression emotion based on TCNN-GRU,and marked the depression classification of the experiment data set.Then we used the TCNN-GRU mode to identify the depression tendency.This paper further proposed the concept of depression index,and constructed the online healthy community depression patient portrait model according to the deep mining of the relationship between the depression index and the degree of depression.The experimental results showed that the TCNN-GRU model can achieve better results in the classification of depression than the traditional convolutional neural network model,recurrent neural network model and hybrid model.Besides,the online healthy community user profile model based on deep learning can also judge the user’s depressive emotion and depressive state from the perspective of text analysis.