Disease Prediction using Symptoms based on Machine Learning Algorithms

Abstract

People are currently suffering from a variety of diseases. Many people are unsure if the symptoms they are experiencing are indicative of a certain disease, and hence they are unable to take the required safeguards. Anticipating the disease during prodromal stage lowers the likelihood of complications. People will not be able to visit a doctor every time they experience a symptom. It may sometimes become a serious ailment if not treated. A model is suggested that uses a variety of symptoms as input to predict the illness. For disease prediction, the suggested method utilizes Decision trees, Naive Bayes, and Random forest classifiers. The ultimate result will be the mode of all these machine learning models. Users will be given a graphical user interface (GUI) to choose their symptoms. The final result will be shown on the interface using all three machine learning techniques, and feature extraction will be done depending on their symptoms. Four modules make up the proposed methodology. Preprocessing will be done on the dataset in the first module. The decision tree classifier is used to generate a prediction model in the second module. The Random forest method is used for forecast the illness in the third module, and the Naive Bayes technique is utilized in the fourth model, with the mode of the outputs from all the three models taken into account.