Depression Detection from Facebook using Machine Learning Techniques

Depression is a chronic ailment and a critical disease that affects day-to-day human activities, feelings, and thoughts. Depression causes sadness or a lack of interest in any day-to-day activities. Today, depression affects more than 200 million individuals worldwide. It may become better if the affected person can avoid the sentiments of desolation, hopelessness, and worthlessness for few days to weeks. To carry out a further research on depression analysis, this research study. As a result, makes use of Facebook data. From the research literature, the proposed research study has observed that there is no lengthy studies have been written for Bengali Facebook. Hence, the proposed research work has compiled a list of Bengali Facebook comments, status updates, and posts. Recently, different supervised learning techniques are used to estimate different kinds of problem and corresponding solutions. The proposed research study has attempted to predict the depression from depression data. Here, the state-of-the-art algorithms like Supporting Vector Machine [SVM], Random Forest, Decision Tree, K-Nearest Neighbors, and Naïve Bayes [NB] are applied and the results are then compared. According to the comparative results, different algorithms deliver similar results and outcomes for all classifiers applied on the proposed research study by analyzing the depression and non-depression data with an accuracy close to 90%.