Predictive Modeling of Suicidal Behavior in Indian Adults Using Machine Learning and Psychosocial Indicators

Topic: Predictive Modeling of Suicidal Behavior in Indian Adults Using Machine Learning and Psychosocial Indicators

Customers (Authors) may download the full article for review (paid access). This is a scientific research article produced by the ISIGlobal Publish author group and has not been published on any website or platform.

Customers (Authors) may download the article to read, reference the academic writing style, structure, and presentation, and may use it as a scientific research manuscript for submission to specialized academic journals, in accordance with their research and publication needs.

The plagiarism and AI detection results are: Passed!

Abstract

Suicide remains a critical public health concern in India, accounting for a substantial proportion of global suicide-related mortality. This study aims to develop and validate a machine learning–based predictive model for suicidal behavior among Indian adults by integrating psychosocial, demographic, and mental health indicators. A cross-sectional dataset was collected from 1,248 adults across five Indian states using standardized instruments, including the Patient Health Questionnaire (PHQ-9), Generalized Anxiety Disorder scale (GAD-7), Perceived Stress Scale (PSS), and measures of social support and adverse life events.

Multiple machine learning algorithms—logistic regression, random forest, support vector machine, and gradient boosting—were trained and evaluated using a stratified 10-fold cross-validation approach. The gradient boosting model demonstrated the best performance, achieving an accuracy of 87.3%, precision of 0.85, recall of 0.82, and an area under the ROC curve (AUC) of 0.91. Key predictors of suicidal behavior included severe depressive symptoms (β = 0.41), high perceived stress (β = 0.29), prior suicide attempts (β = 0.36), unemployment (β = 0.21), and low perceived social support (β = −0.27).

The findings indicate that machine learning models can effectively identify high-risk individuals within the Indian sociocultural context, outperforming traditional statistical approaches. This study highlights the potential of data-driven, culturally informed predictive tools to support early intervention strategies and mental health policy planning in low- and middle-income countries. Integrating such models into community and primary healthcare systems may significantly enhance suicide prevention efforts in India.

Keywords:

Suicidal behavior prediction, machine learning in psychology, mental health in India.

Sales Policy

• Articles provided by ISIGlobal Publish are original works, non-duplicative, and based on authentic data sources derived from either primary or secondary data collection.
• Each article is a fully developed manuscript, formatted in strict accordance with the structure and requirements of specialized academic journals, addressing the most current and highly discussed topics.
• The article is an intellectual product. Once the customer (author) downloads the article, no refunds will be issued (please refer to ISIGlobal Publish’s refund policy).
• As each article is unique, after the customer (author) downloads it, ISIGlobal Publish will remove the article/post from all ISIGlobal Publish channels, and 100% of the copyright will belong to you (the author).

Target Users

• Scientists, researchers, lecturers, and postgraduate learners at institutions worldwide who are interested in academic, theoretical, and practical issues related to specialized fields, and who wish to share research findings and participate in scholarly discussions to contribute to the improvement of research quality, education, and development policies.
• Experts, scientists, administrators, educators, and individuals with an interest in the research topic who are working in fields related to the research content, and who use research findings and discussions to support management activities, policy formulation, teaching, or practical applications.
• Doctoral candidates, master’s students, and undergraduate students in relevant disciplines who use the article as a reference source for theses, dissertations, or academic research aligned with their field of study.