Mental health plays a significant role in our society, as approximately 1 in 5 adults in the United States experience some sort of mental illness each year. Essentially, our application is a pocket therapist. For patients in rural areas, psychiatrists and therapists are often a rarity, and the commute to appointments are often too distant that these patients miss and cancel their appointments. Therabot interacts with patients and allows them to speak to a reliable outlet who can provide comforting messages in times of stress, trauma, when there is simply no one else.

We leveraged machine learning technologies like Stanford Core NLP to run sentiment analysis on text, and combined it with a crf classifier to tag the causes of these sentiments. By analyzing the causes, we could tailor the program's responses such that it matches the user's initial points of complaint or stress.

Share this project:

Updates