The novid-20 is an innovative solution to provide you an online doctor whenever and wherever you need. It was developed to reduce stress and havoc amongst the citizens as seasonal changes and pollution lead to the onset of seasonal flu and allergies which can show symptoms similar to that of COVID-19. Now the fear of the global pandemic instills a doubt of being infected and people frequently feel the urge to visit a hospital. An online solution like this will reduce crowding at hospitals and clinics moreover assuring people about their health and well-being.

Project Development

Our solution for the ongoing pandemic NOvid-20 is a web-based application with its front-end developed using HTML, CSS, and JavaScript.The database for the backend was created using Firebase. The prediction model is developed using Machine Learning with Python and the libraries used for the same are flask,sci-kit-learn,scipy, and gunicorn. The model was deployed on Heroku. Also, we made a chatbot to increase the productivity of our website by using Dialogflow and kommunicate.

The Working

Our website has a prediction model which will accept the responses of the user in the form of 0s and 1s for list symptoms indicating the presence or absence of the symptoms and the result of the input is informed to the user in the form of the severity of symptoms(whether the condition is severe or not). The chatbot will be beneficial in solving queries and doubts of our users to keep them informed.

The Challenges

The process of creation of NOvid-20 had a few challenges in terms of allowing users to check and update their records. Also, the major challenge in this project was the accuracy of the machine learning model used for prediction. Logistic regression was used to predict the outcome with an accuracy of 75% and it was further trained with the dataset. Moreover, our chatbot is unable to process all kinds of user input and we are working over it to improve its functioning.

Future Prospects

With our new learning, we will try to improve the efficiency and user interactivity of our website. Our aim is to improve scalability, make the website platform-independent, and improve the frontend to make it more user friendly. Further, we would like to include the clustering of data to make the results much more informative for our users.

Share this project: