【原】11月3日論文推薦（附下載地址）

學(xué)術(shù)頭條 2020-11-27

展開全文

論文名：
Cross-Domain Depression Detection via Harvesting Social Media

來源：IJCAI/2018

作者：Tiancheng Shen,Jia Jia, Guangyao Shen, Fuli Feng, Xiangnan He,

Huanbo Luan, Jie Tang, Thanassis Tiropanis, Tat-Seng Chua and Wendy Hall

推薦理由：

該論文通過社交網(wǎng)絡(luò)例如Twitter、微博等社交網(wǎng)絡(luò)平臺(tái)的數(shù)據(jù)探測(cè)抑郁癥人群。該論文首先系統(tǒng)地分析了跨領(lǐng)域的抑郁癥相關(guān)特征模式，并總結(jié)了兩個(gè)主要的檢測(cè)挑戰(zhàn)，即異構(gòu)性和發(fā)散性。同時(shí)提出了一個(gè)帶有特征自適應(yīng)變換和組合策略的交叉領(lǐng)域的深度神經(jīng)網(wǎng)絡(luò)模型——DNN-FATC，來實(shí)現(xiàn)跨異構(gòu)域傳遞相關(guān)信息。實(shí)驗(yàn)顯示，與現(xiàn)有的異構(gòu)遷移方法或直接在目標(biāo)域進(jìn)行訓(xùn)練相比該模型方法F1值提高了3.4%以上。

Abstract

Depression detection is a significant issue for human well-being. In previous studies, online detection has proven effective in Twitter, enabling proactive care for depressed users. Owing to cultural differences, replicating the method to other social media platforms, such as Chinese Weibo, however, might lead to poor performance because of insufficient available labeled (self-reported depression) data for model training. In this paper, we study an interesting but challenging problem of enhancing detection in a certain target domain (e.g. Weibo) with ample Twitter data as the source domain. We first systematically analyze the depressionrelated feature patterns across domains and summarize two major detection challenges, namely isomerism and divergency. We further propose a crossdomain Deep Neural Network model with Feature Adaptive Transformation & Combination strategy (DNN-FATC) that transfers the relevant information across heterogeneous domains. Experiments demonstrate improved performance compared to existing heterogeneous transfer methods or training directly in the target domain (over 3.4% improvement in F1), indicating the potential of our model to enable depression detection via social media for more countries with different cultural settings.