Machine Learning-Based Approaches in the Detection of Suicide From Social Media Comments

Machine Learning-Based Approaches in the Detection of Suicide From Social Media Comments

Dipanwita Ghosh, Mihir Sing, Arpan Adhikary, Asit Kumar Nayek
Copyright: © 2023 |Pages: 14
DOI: 10.4018/978-1-6684-7561-4.ch007
OnDemand:
(Individual Chapters)
Available
$37.50
Current Special Offers
No Current Special Offers
TOTAL SAVINGS: $37.50

Abstract

Suicidal tendencies have increased today due to nuclear organization of families and rapid urbanization around the world. Loneliness, aggression, and fast-moving daily lives make the youths and the aged persons depressed. Most of the time, they are involved in mutual relationships on social media. Social media posts and chats, thus, become an important resource from where we can find one's mental illness level and suicidal tendances. The most-used keywords are taken from an open database and are analyzed. ML algorithms like random forest, support vector classifier, and KNN are used to train and predict a person's suicide attempt. Out of these algorithms, SVC produces greater accuracy. To generate more accuracy, word sets shall be robust.
Chapter Preview
Top

In recent past, many researchers have taken ML and DL classifiers to classify suicidal tendency using different comments, posts, chats, etc. from different social media platforms. Mostly used phrases, words, etc. used by those victims are taken care of while we try to analyze. Mostly they are collected from reference literature. In the references, methods of analysis are different from each other and few of them are discussed in this section.

Sharma et al. (2011) used machine learning algorithms to predict suicide risk in individuals using a dataset of over 3,000 patients. The authors compared the performance of several algorithms and found that random forests outperform other models.

Complete Chapter List

Search this Book:
Reset