A Text-Based Approach for Diagnosing Depression Using Social Media Texts

Depression is a mood health issue that can result in persistent feelings of melancholy, a lack of interest in once-enjoyable activities, and physical changes. While many other approaches, like watching facial expressions on video, listening to spoken emotions, reading written responses from the text, and measuring physiology with wearable technology have all been used in past times, multi-modal approaches have also produced encouraging results. This proposed research presents a unique approach for diagnosing depression using social media text data that combines the GloVe word embedding technology with a Long Short-Term Memory (LSTM) neural network model. The method takes advantage of both methodologies strengths to represent social media text data as high-dimensional vectors and capture the temporal connections between words in the text. The work trained a GloVe model on a huge data-set of depression-related social site messages to develop high dimensional numerical vectors to the words. These embedding were then fed into an LSTM model, which classified the text as either depressive or non-depressive. The proposed method was tested on a data-set collected from reddit platform and achieved a classification accuracy of 95%. The findings show that the proposed approach for early diagnosis of depression using social media data is successful, and that it has real-world implications in mental health monitoring and intervention.