Med 277 Project A google assistant bot that aims to predict diseases based on symptoms.
DialogFlow has been used to converse with a patient and a combination of rake, naive bayes classifier and a kaggle dataset, we suggest probable symtoms and diseases to the patient.
The data from the kaggle challenge available here has been used for the predictive tasks. We use dia_3.csv, sym_3.csv and sym_dis_matrix.csv.
You can deploy the server code on heroku from here:
The zip of Dr.Whatson's Dialogflow agents and intents is available here.
To recreate the app:
- Fork/make a copy of the repository. Upload datafiles from the link to a new
datafolder - Create and launch the Heroku server (Click the button above)
- Create a new DialogFlow agent here
- Import all contents of the zip downloaded above into this agent
- Set fulfilment webhook url as
<your heroku-app>/webhook
You will be able to access the app through the google assitant on your devices.