Disease Prediction Based on Symptoms Using Various Machine Learning Techniques

Abstract

Meticulous and prompt analysis of any health related issues is significant for the anticipation and treatment of the disease. The conventional method of determination may not be adequate on account of a genuine infirmity. Fostering a clinical determination framework dependent on machine learning calculations for forecast of any illness can help in a more exact finding than the regular strategy. We have built a disease predication framework utilizing numerous machine learning techniques from symptoms. The dataset utilized had more than 261 illnesses and 500 symptoms for handling. The Random Forest Classifier gave the best outcomes when contrasted with Multinomial Naïve Bayes Classifier, K-Nearest Neighbors, Logistic Regression, Support Vector Machines, Decision Tree, and Multilayer Perceptron Classifier models. The accuracy of the proposed Random Forest Classifier model on the given dataset was 91.06%. Our prediction model can go about as a specialist for the early finding of disease to guarantee the treatment can happen on schedule and lives can be saved.