Mental Health Detection with TF-IDF Feature Extraction

The extensive use of social media, particularly Twitter, provides a data source for post-analysis on the platform. Many writings describe the condition as if a person is under stress. Stress can happen to anyone, and prolonged stress can lead to mental health problems. Text analysis is a method to extract information derived from unstructured data, namely text data. In doing their job, computers cannot process data in the form of text, so the weighting of a word becomes very important, where it can represent words in numerical form. One method of reproducing text data into numerical form is Term Frequency Inverse Document Frequency (TF-IDF). This study aims to analyze text data on mental health using TF- IDF (Term Frequency-Inverse Document Frequency) feature extraction with three learning models for classification, namely SVM (Support Vector Machine), Random Forest, and Decision Tree. The classification results show that SVM has an accuracy rate of 89%, Random Forest 88%, and Decision Tree 83%. The test results show that SVM has better performance than other classification models.