Mental illness detection through harvesting social media: a comprehensive literature review DOI Creative Commons
Shahid Munir Shah, Mahmoud Aljawarneh, Muhammad Aamer Saleem

et al.

PeerJ Computer Science, Journal Year: 2024, Volume and Issue: 10, P. e2296 - e2296

Published: Oct. 7, 2024

Mental illness is a common disease that at its extremes leads to personal and societal suffering. A complicated multi-factorial disease, mental influenced by number of socioeconomic clinical factors, including individual risk factors. Traditionally, approaches relying on interviews filling out questionnaires have been employed diagnose illness; however, these manual procedures found be frequently prone errors unable reliably identify individuals with illness. Fortunately, people illnesses express their ailments social media, making it possible more precisely harvesting media posts. This study offers thorough analysis how (more specifically, depression) from users’ data. Along the explanation data acquisition, preprocessing, feature extraction, classification techniques, most recent published literature presented give readers understanding subject. Since, in past, majority relevant scientific community has focused using machine learning (ML) deep (DL) models illness, so review also focuses techniques along detail, critical presented. More than 100 DL, ML, natural language processing (NLP) based developed for past reviewed, technical contributions strengths are discussed. There exist multiple studies, discussing extensive complete road map design detection system ML DL methods limited. The includes detail dataset may acquired platforms, preprocessed, features extracted employ detection. Hence, we anticipate this will help learn them comprehensive identifying

Language: Английский

Enhancing Arabic text-to-speech synthesis for emotional expression in visually impaired individuals using the artificial hummingbird and hybrid deep learning model DOI Creative Commons

Mahmoud Selim,

Mohammed Assiri

Alexandria Engineering Journal, Journal Year: 2025, Volume and Issue: 119, P. 493 - 502

Published: Feb. 8, 2025

Language: Английский

Citations

2
Deep learning in medicine: advancing healthcare with intelligent solutions and the future of holography imaging in early diagnosis DOI
Asifa Nazir, Ahsan Hussain, Mandeep Singh

et al.

Multimedia Tools and Applications, Journal Year: 2024, Volume and Issue: unknown

Published: July 5, 2024

Language: Английский

Citations

8
Advancing depression detection on social media platforms through fine-tuned large language models DOI
Shahid Munir Shah,

Syeda Anshrah Gillani,

Mirza Baig

et al.

Online Social Networks and Media, Journal Year: 2025, Volume and Issue: 46, P. 100311 - 100311

Published: March 22, 2025

Language: Английский

Citations

1
DepressionEmo: A novel dataset for multilabel classification of depression emotions DOI

Abu Bakar Siddiqur Rahman,

Hoang-Thang Ta,

Lotfollah Najjar

et al.

Journal of Affective Disorders, Journal Year: 2024, Volume and Issue: 366, P. 445 - 458

Published: Aug. 28, 2024

Language: Английский

Citations

5
Utilizing Large Language Models to Detect Depression from User-Generated Diary text data: A Novel Approach in Digital Mental Health Screening (Preprint) DOI Creative Commons
Daun Shin,

Hyoseung Kim,

Seunghwan Lee

et al.

Journal of Medical Internet Research, Journal Year: 2024, Volume and Issue: 26, P. e54617 - e54617

Published: Aug. 11, 2024

Background Depressive disorders have substantial global implications, leading to various social consequences, including decreased occupational productivity and a high disability burden. Early detection intervention for clinically significant depression gained attention; however, the existing screening tools, such as Center Epidemiologic Studies Depression Scale, limitations in objectivity accuracy. Therefore, researchers are identifying objective indicators of depression, image analysis, blood biomarkers, ecological momentary assessments (EMAs). Among EMAs, user-generated text data, particularly from diary writing, emerged analyzable source detecting or diagnosing leveraging advancements large language models ChatGPT. Objective We aimed detect based on through an emotional writing app using model (LLM). validate value semistructured data EMA source. Methods Participants were assessed Patient Health Questionnaire suicide risk was evaluated Beck Scale Suicide Ideation before starting after completing 2-week period. The daily diaries also used analysis. performance LLMs, ChatGPT with GPT-3.5 GPT-4, without fine-tuning training set. comparison involved use chain-of-thought zero-shot prompting analyze structure content. Results 428 91 participants; demonstrated superior detection, achieving accuracy 0.902 specificity 0.955. However, balanced highest (0.844) prompt techniques; it displayed recall 0.929. Conclusions Both GPT-4.0 relatively reasonable recognizing diaries. Our findings highlight potential clinical usefulness depression. In addition measurable indicators, step count physical activity, future research should increasingly emphasize qualitative digital expression.

Language: Английский

Citations

4
Detection of Depression in Social Media Posts using Emotional Intensity Analysis DOI Open Access

M. Kiran Myee,

R. Deepthi Crestose Rebekah,

T K Deepa

et al.

Engineering Technology & Applied Science Research, Journal Year: 2024, Volume and Issue: 14(5), P. 16207 - 16211

Published: Oct. 9, 2024

Tapping into digital footprints on social media, this research focuses providing new insights detecting depression through textual analysis. Initially, emotional raw data found in media posts, aimed particularly at the expressions of anger, fear, joy, and sadness, were collected analyzed. These emotions, each scored by their intensity, offer a quantifiable view users' mental state, serving as possible markers. Central to methodological framework adopted is binary classification system, which classifies texts depressive or non-depressive states, well founded patterns unearthed from data. The proposed model rigorously trains Artificial Intelligence/Machine Learing (AI/ML) models traverse complexities natural language, concentrating noticing delicate indications that signal depression. introduced are tested measured with accuracy, precision, recall, F1-score. RoBERTa, DistilBERT, Electra transformer-based emphasized research. Their performance critically evaluated, results denoting particular capabilities understanding contextualizing key advantage early identification health issues. This stands intersection technology health, revolutionizing monitoring intervention.

Language: Английский

Citations

3
Physics-Constrained Three-Dimensional Swin Transformer for Gravity Data Inversion DOI Creative Commons
Ping Yu,

Longran Zhou,

Shuai Zhou

et al.

Remote Sensing, Journal Year: 2025, Volume and Issue: 17(1), P. 113 - 113

Published: Jan. 1, 2025

This paper proposes a physics-constrained 3D Swin Transformer (ST) for gravity inversion. By leveraging the self-attention mechanism in ST, method effectively models global dependencies within data, enabling network to reweight features globally and focus on critical anomalous regions. Additionally, prior gradient information is integrated into loss function, hierarchical weight allocation strategy adopted guide model learning boundary of density structures deep-seated more effectively. Synthetic experiments demonstrate that proposed achieves lower errors, better alignment, higher inversion accuracy. The approach further validated using anomaly observations from Gonghe Basin Qinghai, yielding reliable precise results.

Language: Английский

Citations

0
Early Depression Detection from Social Media: State-of-the-Art Approaches DOI

A. Alsaedi,

Wael M. S. Yafooz

Studies in computational intelligence, Journal Year: 2025, Volume and Issue: unknown, P. 61 - 75

Published: Jan. 1, 2025

Language: Английский

Citations

0
Sentiment analysis applications using deep learning advancements in social networks: A systematic review DOI
Erfan Bakhtiari Ramezani

Neurocomputing, Journal Year: 2025, Volume and Issue: unknown, P. 129862 - 129862

Published: March 1, 2025

Language: Английский

Citations

0
Text-Based Depression Prediction on Social Media Using Machine Learning: Systematic Review and Meta-Analysis DOI Creative Commons
Doreen Phiri, Frank Makowa, Vivi Leona Amelia

et al.

Journal of Medical Internet Research, Journal Year: 2025, Volume and Issue: 27, P. e59002 - e59002

Published: April 11, 2025

Background Depression affects more than 350 million people globally. Traditional diagnostic methods have limitations. Analyzing textual data from social media provides new insights into predicting depression using machine learning. However, there is a lack of comprehensive reviews in this area, which necessitates further research. Objective This review aims to assess the effectiveness user-generated texts and evaluate influence demographic, language, activity, temporal features on through Methods We searched studies 11 databases (CINHAL [through EBSCOhost], PubMed, Scopus, Ovid MEDLINE, Embase, PubPsych, Cochrane Library, Web Science, ProQuest, IEEE Explore, ACM digital library) January 2008 August 2023. included that used texts, learning, reported area under curve, Pearson r, specificity sensitivity (or for their calculation) predict depression. Protocol papers not written English were excluded. extracted study characteristics, population outcome measures, prediction factors each study. A random effects model was extract effect sizes with 95% CIs. Study heterogeneity evaluated forest plots P values Cochran Q test. Moderator analysis performed identify sources heterogeneity. Results total 36 included. observed significant overall correlation between depression, large size (r=0.630, CI 0.565-0.686). noted same demographic (largest size; r=0.642, 0.489-0.757), activity (r=0.552, 0.418-0.663), language (r=0.545, 0.441-0.649), (r=0.531, 0.320-0.693). The platform type (public or private; P<.001), learning approach (shallow deep; P=.048), use measures (yes no; P<.001) moderators. Sensitivity revealed no change results, indicating result stability. Begg-Mazumdar rank (Kendall τb=0.22063; P=.058) Egger test (2-tailed t34=1.28696; P=.207) confirmed absence publication bias. Conclusions Social content can be useful tool Demographics, should considered maximize accuracy models. Additionally, type, approach, models need attention. challenging, findings may apply broader population. Nevertheless, our offer valuable future Trial Registration PROSPERO CRD42023427707; https://www.crd.york.ac.uk/PROSPERO/view/CRD42023427707

Language: Английский

Citations

0