Characterizing Suicide Ideation by Using Mental Disorder Features on Microblogs: A Machine Learning Perspective

Samer Sarsam; Hosam Al-Samarraie; Ahmed Ibrahim Alzahrani; Chit Su  Mon; Abdul Samad  Shibghatullah

doi:10.1007/s11469-022-00958-z

Characterizing Suicide Ideation by Using Mental Disorder Features on Microblogs: A Machine Learning Perspective

Samer Sarsam, Hosam Al-Samarraie, Ahmed Ibrahim Alzahrani, Chit Su Mon, Abdul Samad Shibghatullah

Research output: Contribution to journal › Article › peer-review

7 Citations (Scopus)

99 Downloads (Pure)

Abstract

Despite the success of psychological and clinical methods, psychological studies revealed that the number of individuals exhibiting suicide ideation has highly increased in the recent decades. This study explored the potential of using certain sentimental features as a means for characterizing suicide. A total of 54,385 English-language tweets were collected and processed to extract suicide-related topics using the Latent Dirichlet Allocation (LDA) algorithm. Both suicidal polarity (positive, negative, and neutral) and emotions (anger, fear, sadness, and trust) were extracted via SentiStrength, time series, and NRC Affect Intensity Lexicon methods. The results showed that suicidal tweets were less associated with trust, anger, and positive sentiments. In contrast, fear, sadness, and negative sentiments were highly associated with suicidal statements. The prediction results using this approach showed 97.64% accuracy in detecting suicide ideation. The obtained results from analyzing suicide-related tweets hold a promising future for characterizing suicide ideation worldwide.

Original language	English
Pages (from-to)	1783–1796
Number of pages	14
Journal	International Journal of Mental Health and Addiction
Volume	22
Issue number	4
Early online date	14 Nov 2022
DOIs	https://doi.org/10.1007/s11469-022-00958-z
Publication status	Published - Aug 2024

Bibliographical note

Copyright © and Moral Rights are retained by the author(s) and/ or other copyright owners. A copy can be downloaded for personal non-commercial research or study, without prior permission or charge. This item cannot be reproduced or quoted extensively from without first obtaining permission in writing from the copyright holder(s). The content must not be changed in any way or sold commercially in any format or medium without the formal permission of the copyright holders.

This document is the author’s post-print version, incorporating any revisions agreed during the peer-review process. Some differences between the published version and this version may remain and you are advised to consult the published version if you wish to cite from it.

Funder

This work was funded by the Researchers Supporting Project number [RSP-2021/157], King Saud University, Riyadh, Saudi Arabia

Funding

This work was funded by the Researchers Supporting Project number [RSP-2021/157], King Saud University, Riyadh, Saudi Arabia.

Funders	Funder number
King Saud University	RSP-2021/157

Keywords

Suicide ideation
Sentiment analysis
Topic modeling
Twitter

UN SDGs

This output contributes to the following UN Sustainable Development Goals (SDGs)

Access to Document

10.1007/s11469-022-00958-z

Post-PrintAccepted author manuscript, 1.44 MB

Cite this

@article{b416deaf6e4b4fd5a98dc5cb8ea0553e,

title = "Characterizing Suicide Ideation by Using Mental Disorder Features on Microblogs: A Machine Learning Perspective",

abstract = "Despite the success of psychological and clinical methods, psychological studies revealed that the number of individuals exhibiting suicide ideation has highly increased in the recent decades. This study explored the potential of using certain sentimental features as a means for characterizing suicide. A total of 54,385 English-language tweets were collected and processed to extract suicide-related topics using the Latent Dirichlet Allocation (LDA) algorithm. Both suicidal polarity (positive, negative, and neutral) and emotions (anger, fear, sadness, and trust) were extracted via SentiStrength, time series, and NRC Affect Intensity Lexicon methods. The results showed that suicidal tweets were less associated with trust, anger, and positive sentiments. In contrast, fear, sadness, and negative sentiments were highly associated with suicidal statements. The prediction results using this approach showed 97.64% accuracy in detecting suicide ideation. The obtained results from analyzing suicide-related tweets hold a promising future for characterizing suicide ideation worldwide.",

keywords = "Suicide ideation, Sentiment analysis, Topic modeling, Twitter",

author = "Samer Sarsam and Hosam Al-Samarraie and Alzahrani, {Ahmed Ibrahim} and Mon, {Chit Su} and Shibghatullah, {Abdul Samad}",

note = "Copyright {\textcopyright} and Moral Rights are retained by the author(s) and/ or other copyright owners. A copy can be downloaded for personal non-commercial research or study, without prior permission or charge. This item cannot be reproduced or quoted extensively from without first obtaining permission in writing from the copyright holder(s). The content must not be changed in any way or sold commercially in any format or medium without the formal permission of the copyright holders. This document is the author{\textquoteright}s post-print version, incorporating any revisions agreed during the peer-review process. Some differences between the published version and this version may remain and you are advised to consult the published version if you wish to cite from it. ",

year = "2024",

month = aug,

doi = "10.1007/s11469-022-00958-z",

language = "English",

volume = "22",

pages = "1783–1796",

journal = "International Journal of Mental Health and Addiction",

issn = "1557-1874",

publisher = "Springer ",

number = "4",

}

TY - JOUR

T1 - Characterizing Suicide Ideation by Using Mental Disorder Features on Microblogs

T2 - A Machine Learning Perspective

AU - Sarsam, Samer

AU - Al-Samarraie, Hosam

AU - Alzahrani, Ahmed Ibrahim

AU - Mon, Chit Su

AU - Shibghatullah, Abdul Samad

N1 - Copyright © and Moral Rights are retained by the author(s) and/ or other copyright owners. A copy can be downloaded for personal non-commercial research or study, without prior permission or charge. This item cannot be reproduced or quoted extensively from without first obtaining permission in writing from the copyright holder(s). The content must not be changed in any way or sold commercially in any format or medium without the formal permission of the copyright holders. This document is the author’s post-print version, incorporating any revisions agreed during the peer-review process. Some differences between the published version and this version may remain and you are advised to consult the published version if you wish to cite from it.

PY - 2024/8

Y1 - 2024/8

N2 - Despite the success of psychological and clinical methods, psychological studies revealed that the number of individuals exhibiting suicide ideation has highly increased in the recent decades. This study explored the potential of using certain sentimental features as a means for characterizing suicide. A total of 54,385 English-language tweets were collected and processed to extract suicide-related topics using the Latent Dirichlet Allocation (LDA) algorithm. Both suicidal polarity (positive, negative, and neutral) and emotions (anger, fear, sadness, and trust) were extracted via SentiStrength, time series, and NRC Affect Intensity Lexicon methods. The results showed that suicidal tweets were less associated with trust, anger, and positive sentiments. In contrast, fear, sadness, and negative sentiments were highly associated with suicidal statements. The prediction results using this approach showed 97.64% accuracy in detecting suicide ideation. The obtained results from analyzing suicide-related tweets hold a promising future for characterizing suicide ideation worldwide.

AB - Despite the success of psychological and clinical methods, psychological studies revealed that the number of individuals exhibiting suicide ideation has highly increased in the recent decades. This study explored the potential of using certain sentimental features as a means for characterizing suicide. A total of 54,385 English-language tweets were collected and processed to extract suicide-related topics using the Latent Dirichlet Allocation (LDA) algorithm. Both suicidal polarity (positive, negative, and neutral) and emotions (anger, fear, sadness, and trust) were extracted via SentiStrength, time series, and NRC Affect Intensity Lexicon methods. The results showed that suicidal tweets were less associated with trust, anger, and positive sentiments. In contrast, fear, sadness, and negative sentiments were highly associated with suicidal statements. The prediction results using this approach showed 97.64% accuracy in detecting suicide ideation. The obtained results from analyzing suicide-related tweets hold a promising future for characterizing suicide ideation worldwide.

KW - Suicide ideation

KW - Sentiment analysis

KW - Topic modeling

KW - Twitter

UR - http://www.scopus.com/inward/record.url?scp=85141951658&partnerID=8YFLogxK

U2 - 10.1007/s11469-022-00958-z

DO - 10.1007/s11469-022-00958-z

M3 - Article

SN - 1557-1874

VL - 22

SP - 1783

EP - 1796

JO - International Journal of Mental Health and Addiction

JF - International Journal of Mental Health and Addiction

IS - 4

ER -

Characterizing Suicide Ideation by Using Mental Disorder Features on Microblogs: A Machine Learning Perspective

Abstract

Bibliographical note

Funder

Funding

Keywords

UN SDGs

Access to Document

Other files and links

Fingerprint

Cite this