As we already know about the taboo regarding the mental health we thought of helping out people in very simple yet effective manner which is very discrete as well.

What it does

We ask the user few questions that provides us with data which lets us understand the emotional psyche of the user. After that our model does emotional analysis which is based on a probabilistic approach and thus provides the user with a simple observation of whether or not they should consult a psychiatrist.

How we built it

We thought of building a web based solution using different ML/DL models. We tried to build a mental health Q&A based chatbot which would answer basic queries regarding mental health. This uses a concept of sentence embeddings and cosine similarity. For generating sentence embeddings we used sentence transformers. For detecting mental health problems we ask a few set of questions and try to perform emotion analysis. On the basis of this we try to detect mental health issues. This emotion analysis is done using our own emotion recognition API. If the person has or if the system detected some mental health issues than it will give the information about nearby practioners.

Challenges we ran into

Dataset availability, Model size and deployment issue.

Accomplishments that we're proud of

Successfully created a prototype that could perform emotional analysis.

What we learned

Deep learning and NLP

What's next for Mental Health App

This app is in a very rudimentary stage and we would like to have a full fledged ecosystem where doctors are also associated and they get a preliminary report of the user beforehand and they can then further asses the user in a very better way with prior knowledge of the person. Along with that

Share this project: