Abstract :

According To The World Health Organization (WHO), Over 970 Million People Globally Were Living With A Mental Disorder In 2019, And Depression Is The Leading Cause Of Disability Worldwide. Despite The Growing Mental Health Crisis, Early Detection And Personalized Intervention Remain Significantly Underexplored, Especially Among Working Individuals Where Stigma And Underreporting Are Common. Traditional Screening Methods Rely On Self-reports Or Clinical Evaluations, Which Are Often Timeconsuming, Inconsistent, And Inaccessible. Additionally, Many Current Machine Learning Models Lack Effective Handling Of Linguistic Nuances In Textual Data And Fail To Integrate Contextual Attributes Such As Lifestyle Or Demographic Factors. This Study Presents A Comprehensive Natural Language Processing (NLP)-based Pipeline For Mental Health Classification That Integrates Both Linguistic Cues And Personal Metadata For Improved Prediction. The Dataset Consists Of Multiple Features Including Raw Textual Responses (text) And Structured Inputs Like Age, Gender, Employment_status, And Depression_score. This Work First Applies NLP Preprocessing Techniques Including Tokenization, Stopword Removal, And Lemmatization To Clean The Text. A Thorough Exploratory Data Analysis (EDA) Uncovers Trends And Correlations Between Mental Health Indicators And Lifestyle Variables Such As Sleep Hours And Stress Levels. TF-IDF Vectorization Is Employed To Transform The Processed Text Into Weighted Numerical Features That Highlight Important Terms Relevant To Mental Health Expression. We Then Train And Evaluate Multiple Classifiers: Decision Tree Classifier (DTC), Random Forest Classifier (RFC), Logistic Regression, And Light Gradient Boosting Machine (LGBM). Among These, LGBM Achieved The Best Performance, With An Accuracy Of 95.97%, Precision Of 96.04%, Recall Of 95.94%, And F1-score Of 95.97%. This High Accuracy Demonstrates The Model’s Strong Ability To Detect Mental Health Risk Based On Linguistic And Cont

Published:

18-8-2025

Issue:

Vol. 25 No. 8 (2025)

Page Nos:

231-245

Section:

Articles

License:

This work is licensed under a Creative Commons Attribution-NonCommercial-NoDerivatives 4.0 International License.

How to Cite