Detecting Suicidal Ideation in Chinese Microblogs with Psychological Lexicons

Abstract

Suicide is among the leading causes of death in China. However, technical approaches toward preventing suicide are challenging and remaining under development. Recently, several actual suicidal cases were preceded by users who posted microblogs with suicidal ideation to Sina Weibo, a Chinese social media network akin to Twitter. It would therefore be desirable to detect suicidal ideations from microblogs in real-time, and immediately alert appropriate support groups, which may lead to successful prevention. In this paper, we propose a real-time suicidal ideation detection system deployed over Weibo, using machine learning and known psychological techniques. Currently, we have identified 53 known suicidal cases who posted suicide notes on Weibo prior to their deaths. We explore linguistic features of these known cases using a psychological lexicon dictionary, and train an effective suicidal Weibo post detection model. 6714 tagged posts and several classifiers are used to verify the model. By combining both machine learning and psychological knowledge, SVM classifier has the best performance of different classifiers, yielding an F-measure of 68.3%, a Precision of 78.9%, and a Recall of 60.3%.

Researching Mental Health Disorders in the Era of Social Media: Systematic Review

“…The techniques used in these studies included support vector machine (SVM) [32,33,35,38,42,56,69], linear SVM [24,27,41,46,60], and SVM with a radial basis function kernel [24,27,46,51,65-67]. Regression techniques included ridge regression [28], linear regression [37,58], log-linear regression [53,59], logistic regression [25,31,33,37,48,49,51], binary logistic regression with elastic net regularization [41,43], linear regression with stepwise selection [39,55,64], stepwise logistic regression with forward selection [50], regularized multinomial logistic regression [29], linear support vector regression [45,55], least absolute shrinkage and selection operator [55,68], and multivariate adaptive regression splines [55]. Other algorithms used for binary classification were decision trees [35,51,56,62,63], random forest [26,48,51], rules decision [62], naive Bayes [24,35,51,56,62,69], k-nearest neighbor [24,56], maximum entropy [42], neural network [69], and deep learning neural network [57].…”

Detection of Suicide Ideation in Social Media Forums Using Deep Learning

“…Desmet et al [20] built a suicide note analysis method to detect suicide ideation using binary Support Vector Machine (SVM) classifiers. Huang et al [21] created a psychological lexicon based on a Chinese sentiment dictionary (Hownet). He applied the SVM approach to identify a classification for developing a real-time suicide ideation detection system deployed in Chinese Weibo.…”

Exploring and Learning Suicidal Ideation Connotations on Social Media with Deep Learning

“…A suicide note analysis method for automated the identification of suicidal ideation was built using binary support vector machine classifiers by Desmet and Hoste (2013) using fine-grained emotion detection for classifier optimization with lexico-semantic features for optimization. In 2014, Huang et al (2014) used rule-based methods with a hand-crafted unsupervised classification for developing a real-time suicidal ideation detection system deployed over Weibo 1 , a microblogging platform. By combining both machine learning and psychological knowledge, they reported an SVM classifier as having the best performance of different classifiers.…”