Comparison of naïve bayes, logistic regression and support vector machine for predicting suicidal tendency from social media content

Abstract

The objective of this study is to employ Machine Learning methodologies in order to forecast the chances of depression and suicide among individuals in Nepal by analyzing their social media engagement. The dataset consisting of 2200 entries was subjected to analysis using the Naïve Bayes, Logistic Regression, and Support Vector Machine techniques. The accuracy of the Support Vector Machine was found to be 95.45%. The timely identification of suicidal and depressive incidents had a crucial role in addressing depression and diminishing suicide rates. This study assesses the sensitivity and accuracy of various Machine Learning algorithms in the context of early and late detection.