Machine Learning Approaches for Detecting Early-Stage Depression using Text

The work in this paper identifies the risk of depression in people based on their response which is in the form of text. This method of analyzing risk of depression is called identifying early stage of depression, which is characterized by various symptoms like loneliness, disinterest, etc. The work focuses on analyzing the sentences collected from the people to detect whether the person is in risk of depression or not. Different Machine Learning (ML) algorithms are used in this work to detect depression. The experimental results that we obtained show that the Random Forest Classifier (RFC) gives better performance when compared to other classifiers. An automated system is developed to take input from people and give predictions based on the response.