言为心声:人工智能探测抑郁症的语言证据与价值  

Voice for Heart:The Linguistic Evidence and Value of AI in Detecting Depression

作　　者：伍麟 彭子剑 WU Lin;PENG Zi-jian(School of Sociology,Wuhan University,Wuhan,Hubei,430072,PRC)

机构地区：[1]武汉大学社会学院,湖北武汉430072

出　　处：《西北师大学报（社会科学版）》2020年第5期136-144,共9页Journal of Northwest Normal University(Social Sciences)

基　　金：国家社会科学基金重大研究专项“加强社会心理服务体系建设研究”(18VZL009)。

摘　　要：由于压力驱使、病耻逃离、社会比较、自我呈现等因素,导致了抑郁症个体表现出回避进行线下社交,而偏向进行线上活动。互联网社交媒体已经成为抑郁症个体进行情绪宣泄和言语表达的一个重要场所。抑郁症个体社交媒体发文具有负面型的文风、格式化的用词、含混式的表达、规律性的时间等特征。随着心理学和计算科学的合作日益增多,学术界越来越流行通过大数据、机器学习、人工智能等技术发现用户语言和心理之间的关联,推测其过度化自我聚焦、偏差式社会比较、压抑型认知图式以及隐含伤感类自我呈现等心理状态,判断其罹患抑郁症的可能性,进而掌握抑郁症人群的社会人口学特征以及开展有效干预。通过分析互联网抑郁症人群的大数据分布等社会指标,对于抑郁症个体的康复、相关部门的有效干预以及国民身心健康发展等都具有重要意义。Due to factors such as pressure drive,stigma escape,social comparison,self-presentation,individuals with depression have shown to avoid offline social activities and prefer online activities.Internet has become an essential place for depressed individuals to express their emotions and ideas.Social media postings of depressed individuals have negative literary styles,formatted wording,ambiguous expressions,regular time and other features.With the increasing collaboration between psychology and computational science,it is becoming increasingly popular for academics to use big data,machine learning,artificial intelligence,and etc.Technology discovers correlations between users’language and psychology,inferring their hyper-focused self,deviant social comparisons,and repressive cognitive schemas.The psychological state of depressed people,such as the self-presentation of feelings and implied sadness,can be used to determine their likelihood of depression and thus to understand the social characteristics of depressed people by demographic characteristics and the development of effective interventions.By analyzing social indicators,such as the distribution of big data of depressed people on the Internet,it is possible to make a significant contribution to the recovery of depressed individuals and to the development of effective interventions in relevant sectors.Interventions,as well as the healthy development of the national body and mind,are of great importance.

关 键 词：抑郁症 语言 互联网社交媒体 大数据预测抑郁症 

分 类 号：R749.4[医药卫生—神经病学与精神病学]