Implementing Sentiment Analysis with Applied Nonlinear Analysis: Predicting Mental Disorders from Diverse Online Social Network Datasets

Abstract

Introduction: WHO projected that the depression disorder will increase over the next two decades; therefore, it should be diagnosed early and the sooner it can be detected and prevented. Several machine learning algorithms that may predict depression will be compared in this paper: CNN, LSTM, Bidirectional LSTM, Naive Bayes, Logistic Regression, Random Forest, AdaBoost, and Support Vector Machine. The results show that the models of deep learning, mainly variants of CNN and LSTM, exhibit high accuracy, and traditional classifiers perform exceptionally well. These results highlight how machine learning can contribute towards developing useful tools for the prevention of mental health disorders early in life. In modern life, online social platforms have very subtly integrated into the daily routine of a significant proportion of global citizens.