THE PREVENTION AND PREDICTION OF SUICIDE AND DEPRESSION: A POPULATION-BASED STUDY ON MACHINE LEARNING APPROACH

Suicide is a rare event with a complex basis; it can be challenging to identify people at risk accurately. In this study, suicide among depressed people was predicted using machine learning approaches. Four machine learning algorithms-Logistic Regression (LR), Naive Bayes (NB), Long Short-Term Memory (LSTM), and Support Vector Machine (SVM) were used to create the predictive models. Due to environmental factors, males frequently have higher rates of suicide and mortality than girls. As the annual suicide rate increases, so does the global suicide rate. The LSTM prediction revealed the maximum observed suicide rates for the issue's year and the associated anticipated rates. Compared to a baseline model, this novel approach improves accuracy and weighted F1 score, obtaining 0.93 accuracies and 0.93 weighted F1 score for the Reddit dataset. Significant suicide risk factors in the general population include both depression and personality traits.