Medical Questioning Chatbot

Case Study

In this Project, we created more than 20 chatbots regarding health / medical questions.
All the chatbot was created for a platform where doctors/students who want to become a doctor can practice how and what type of questions should be asked to patient.
This chatbot helps them to cover almost a variety of questions with more than 15 categories where doctors can practice asking questions.
Moreover, there will be an analysis of the chat will also be generated which helps the user to understand where and what kind of questions he/she missed and where he/she got the highest score.
There is also an advanced feature of voice-to-text and text-to-voice on the website so if any wants to access the chatbot without typing questions he/she can easily achieve that by using this advanced feature.

Technologies used

We created all the chatbots in python using NLP.
The machine learning model was implemented in Tensorflow and Tensorflow Lite.
Moreover, for grammar extraction, we used the NLTK library of python.
To implement text-to-voice and voice-to-text, we used Google’s library available in python for converting given voice data into text format and then pass to the model for prediction of output.

Difficulties we faced

The most challenging part of this project was that there were 20 chatbots that were running at the same time to the time and the processing power consumed was way too high.

Solutions

To overcome this difficulty we used Tensorflow Lite which converts the heavy Tensorflow model into a lite weight model which will require low processing power compared to the large model of Tensorflow.