Depressive and Non-depressive Tweets Classification using a Sequential Deep Learning Model

Abstract

Nowadays, depression is one of the most predominant mental health ailments, even causing demise of many lives, especially among the youngsters and middle-aged persons. Use of social media and posting individual expression on it is very much use to for the people. So, social media posts are very useful for detecting a person is in depression or not. In this paper, we have considered depressive and non-depressive tweets classification. We have proposed a sequential Deep learning model with three layers: Embedded, lD-Convolutional and LSTM Layer. The performance of the proposed model has been compared with three traditional Machine Learning models Naïve Bayes, KNN, Random Forest for depressive and non-depressive tweets classification. Our proposed approach of sequential Deep Learning model has achieved 98.47% accuracy, which is better than other models we have compared with.