Predicting Suicide Risk using Machine Learning Techniques

This project aims to predict suicide risk using machine learning techniques on post-mortem data that includes both clinical and demographic information. The process consists of the following steps: Data Pre-processing, Exploratory Data Analysis (EDA), Feature Extraction, Feature Importance Analysis, Model Training, Model Improvement, and Model Performance Analysis. In the Feature Extraction phase, 497 features were extracted from the text data, reducing the number of features from 28,000 to 497 without any loss of information. The Feature Importance Analysis was conducted to identify the most crucial features for the classification task. Logistic Regression, Decision Tree, Support Vector Machine, Random Forest, and Deep Learning models were applied and compared.