Health-Web-App (Website Link)
There is a huge demand for healthcare services worldwide and many countries are experiencing a shortage of physicians. It is believed that AI in healthcare can bring major improvements within the healthcare ecosystem. Building a web portal that can monitor the prediction of health diseases can be beneficial for many people to save time by predicting the presence of a disease on their own. These automated services can help physicians in diagnosing diseases.
I have created a website that helps patients diagnose the disease by entering their symptoms. I analyzed the datasets for the diseases and preprocessed the datasets that can be used in the model building. I implemented various machine learning algorithms and deep learning algorithms on the datasets.
I optimized various algorithms to achieve the high accuracy of prediction with the datasets. A machine learning model with the highest accuracy was used to embed in the Flask application.
One feature that the web system provides is the threshold selection by the user to predict the prediction of the disease. This way the user is not bound by the default 50% as the threshold for the disease prediction. This application is also deployed on the Heroku cloud which can be beneficial for the users globally to use the application.
- Tools used: Python, Machine Learning, Flask, Heroku
- Machine Learning Algorithms used: Logistic Regression, Naive Bayes, XG-Boost, Random Forest, Fully Connected Neural networks
