Constructing Depression Prediction Model using Machine Learning Algorithms

Depression is a prevalent mental health disorder with significant impacts on individuals and society. Given the increasing presence of depressive expressions on social media, this study constructs a predictive model to classify depression-related content using machine learning algorithms. Six algorithms for comparing a model, such as Decision Trees, Random Forest, K-nearest neighbors, Support Vector Machines (SVM), Logistic Regression, and Convolutional Neural Networks (CNNs), are used to identify the most effective model for accurately predicting depressive posts. The study utilized a dataset from Twitter, balancing depressive and non-depressive texts for unbiased training. Text preprocessing involved tokenization, stop word removal, and feature extraction using TF-IDF. Evaluation metrics such as accuracy, precision, recall, and F1-score were applied to assess performance. The results showed that Decision Trees and Logistic Regression performed best, with accuracy exceeding 99%. These models were implemented in a prototype for real-time depression risk prediction. Future work aims to expand the model's applicability to Thai language texts, enhancing support for mental health care