Mental Health Status Prediction Using ML Classifiers with NLP-Based Approaches

Nowadays, people use different online platforms for their day-to-day communication. It is possible to use Natural language processing approaches to infer people's mental states from what they share on these platforms via text or speech. Predicting mental health problems by SMS is a proactive step toward improved treatment. NLP is changing the way mental health experts evaluate the free expression designed to check and diagnose mental diseases in patients. Using machine learning techniques can provide new ways to study human attitudes and behaviors as well as identifying signs and symptoms of mental illness. In this paper, we used NLP for detecting the status of mental health through a text message and our research gives a detailed investigation of several supervised classifier algorithms. Peoples suffer from different types of mental illnesses, but bipolar, panic disorder, depression, stress, anxiety, and PTSD are popular. For this analysis, We classified data using Logistic Regression, BernoulliNB, K-Nearest Neighbors, Random Forest, and Decision Trees. In our proposed method, Logistic Regression shows the best performance compared with the other four classifiers. The experimental outcome supports the proposed methodology's ability to classify patient data more accurately. The proposed model was shown to be effective with an accuracy rate of 93 percent.