Zhengping Jiang, Sarah Ita Levitan, Jonathan Zomick, Julia Hirschberg
{"title":"基于深度情境化表征的Reddit用户心理健康检测","authors":"Zhengping Jiang, Sarah Ita Levitan, Jonathan Zomick, Julia Hirschberg","doi":"10.18653/v1/2020.louhi-1.16","DOIUrl":null,"url":null,"abstract":"We address the problem of automatic detection of psychiatric disorders from the linguistic content of social media posts. We build a large scale dataset of Reddit posts from users with eight disorders and a control user group. We extract and analyze linguistic characteristics of posts and identify differences between diagnostic groups. We build strong classification models based on deep contextualized word representations and show that they outperform previously applied statistical models with simple linguistic features by large margins. We compare user-level and post-level classification performance, as well as an ensembled multiclass model.","PeriodicalId":448872,"journal":{"name":"International Workshop on Health Text Mining and Information Analysis","volume":"109 1","pages":"0"},"PeriodicalIF":0.0000,"publicationDate":"2020-11-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":"44","resultStr":"{\"title\":\"Detection of Mental Health from Reddit via Deep Contextualized Representations\",\"authors\":\"Zhengping Jiang, Sarah Ita Levitan, Jonathan Zomick, Julia Hirschberg\",\"doi\":\"10.18653/v1/2020.louhi-1.16\",\"DOIUrl\":null,\"url\":null,\"abstract\":\"We address the problem of automatic detection of psychiatric disorders from the linguistic content of social media posts. We build a large scale dataset of Reddit posts from users with eight disorders and a control user group. We extract and analyze linguistic characteristics of posts and identify differences between diagnostic groups. We build strong classification models based on deep contextualized word representations and show that they outperform previously applied statistical models with simple linguistic features by large margins. We compare user-level and post-level classification performance, as well as an ensembled multiclass model.\",\"PeriodicalId\":448872,\"journal\":{\"name\":\"International Workshop on Health Text Mining and Information Analysis\",\"volume\":\"109 1\",\"pages\":\"0\"},\"PeriodicalIF\":0.0000,\"publicationDate\":\"2020-11-01\",\"publicationTypes\":\"Journal Article\",\"fieldsOfStudy\":null,\"isOpenAccess\":false,\"openAccessPdf\":\"\",\"citationCount\":\"44\",\"resultStr\":null,\"platform\":\"Semanticscholar\",\"paperid\":null,\"PeriodicalName\":\"International Workshop on Health Text Mining and Information Analysis\",\"FirstCategoryId\":\"1085\",\"ListUrlMain\":\"https://doi.org/10.18653/v1/2020.louhi-1.16\",\"RegionNum\":0,\"RegionCategory\":null,\"ArticlePicture\":[],\"TitleCN\":null,\"AbstractTextCN\":null,\"PMCID\":null,\"EPubDate\":\"\",\"PubModel\":\"\",\"JCR\":\"\",\"JCRName\":\"\",\"Score\":null,\"Total\":0}","platform":"Semanticscholar","paperid":null,"PeriodicalName":"International Workshop on Health Text Mining and Information Analysis","FirstCategoryId":"1085","ListUrlMain":"https://doi.org/10.18653/v1/2020.louhi-1.16","RegionNum":0,"RegionCategory":null,"ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":null,"EPubDate":"","PubModel":"","JCR":"","JCRName":"","Score":null,"Total":0}
Detection of Mental Health from Reddit via Deep Contextualized Representations
We address the problem of automatic detection of psychiatric disorders from the linguistic content of social media posts. We build a large scale dataset of Reddit posts from users with eight disorders and a control user group. We extract and analyze linguistic characteristics of posts and identify differences between diagnostic groups. We build strong classification models based on deep contextualized word representations and show that they outperform previously applied statistical models with simple linguistic features by large margins. We compare user-level and post-level classification performance, as well as an ensembled multiclass model.