Archives

Disease Prediction, Drugs Suggestion and Physician Recommendation System Using Supervised Machine Learning


V. Gireesh, S. Vimal and A.S. Jeyakrishna
Abstract

The primary objective of this work is to provide virtual medical assistance and awareness, second opinions and trustworthy methods of disease prediction through our implementation of a model that would predict diseases from symptoms using Machine Learning Techniques and suggest the appropriate precautions, drugs and physicians for appointment. Most of the similar systems focus either on specific diseases (heart, cancer, diabetes) and feedback systems (exclusively prediction or physician appointment systems), and very few can do a combination of both effectively. In models where the accuracy of prediction is really high, there are either very few options for medical assistance, or the services are exorbitantly priced. We attempt to resolve all of this with our proposed model - the users are asked to select five of their most prominent symptoms. After this, three algorithms - KNN, Random Forest and Gaussian Naïve Bayes are used. Three diseases are predicted using each of the three techniques, along with the accuracy of prediction of the same. Precautions for the predicted diseases are suggested, along with the recommended drugs and side effects. Taking the location of the user as the input, the details of the physicians in the surrounding areas are listed. Accuracy scores and performance visualizations for the diseases are displayed in the console window.

Volume 11 | 03-Special Issue

Pages: 1491-1512