Detecting Suicidal Ideation in Social Media: An Ensemble Method Based on Feature Fusion

Abstract

Suicide has become a serious problem, and how to prevent suicide has become a very important research topic. Social media provides an ideal platform for monitoring suicidal ideation. This paper presents an integrated model for multidimensional information fusion. By integrating the best classification models determined by single and multiple features, different feature information is combined to better identify suicidal posts in online social media. This approach was assessed with a dataset formed from 40,222 posts annotated by Weibo. By integrating the best classification model of single features and multidimensional features, the proposed model ((BSC + RFS)-fs, WEC-fs) achieved 80.61% accuracy and a 79.20% F1-score. Other representative text information representation methods and demographic factors related to suicide may also be important predictors of suicide, which were not considered in this study. To the best of our knowledge, this is the good try that feature combination and ensemble algorithms have been fused to detect user-generated content with suicidal ideation. The findings suggest that feature combinations do not always work well, and that an appropriate combination strategy can make classification models work better. There are differences in the information contained in different functional carriers, and a targeted choice classification model may improve the detection rate of suicidal ideation.

The effects of short video app-guided loving-kindness meditation on college students’ mindfulness, self-compassion, positive psychological capital, and suicide ideation

“…We will send these compliance data to the researchers every 2 weeks for the duration of the study. Taking advantage of this information, the researchers calculated and estimated the total effective time of each participant using the short video application of LKM during the 8-week intervention period (Chen et al, 2022 , 2021a , 2021b ; Liu et al, 2022a , 2022b , 2022c , 2022d , 2022e , 2022f ; Liu et al, 2022a , 2022b , 2022c , 2022d , 2022e , 2022f ).…”

“…It refers to death, suicide, and serious self-harm in an individual’s thoughts or thoughts, including thoughts about plans, steps, and results of suicide (Zortea et al, 2020 ). Compared with traditional media, the Internet and social networking sites that are spread described that suicide and suicide knowledge provides more opportunities, according to a study of Chinese students, suicidal ideation and suicide-related highly relevant social media use behavior, including suicide information, reviews, and forward to commit suicide, or talk about suicide-related content (Liu et al, 2022a , 2022b , 2022c , 2022d , 2022e , 2022f ). Although the vast majority of suicidal ideations do not lead to actual suicidal actions, suicidal ideation is still considered a necessary condition for suicide and attempted suicide.…”

“…Positive psychological capital enhances an individual’s psychological resilience. When faced with life’s stresses and setbacks, those with a high level of positive psychological capital are more likely to persevere, seek solutions to problems, and refrain from choosing suicide as an option (Liu et al, 2022a , 2022b , 2022c , 2022d , 2022e , 2022f ). Their optimistic outlook and sense of hope can serve as powerful motivators to overcome adversity.…”

“…LKM, a method of work that concentrates on blessing oneself and others in the mind, has the effect of generating positive emotions and fostering positive attitudes (Chen et al, 2021a , 2021b ; Liu et al, 2020a , 2020b , 2020c , 2022a , 2022b , 2022c , 2022d , 2022e , 2022f ). As for the function and effect of LKM, several intervention strategies based on LKM have been developed in the academic circle, mainly focusing on two aspects positive attitude and positive emotion.…”

A Hybrid Attention Based Deep Learning System for Suicidal Ideation Detection in Social Media

“…First, the MA model transforms , , and into sub-spaces, with different, learnable linear projections. It is expressed as follows: (11) Where, indicates the head of query, key, and value, respectively and refers to the weight matrices of query, key and value. Furthermore, attention functions are applied simultaneously to get the output features and it is mathematically expressed as follows:…”