Disease Prediction Based on Symptoms Using Ensemble and Hybrid Machine Learning Models

The health sector is one of the prominent sectors in which continuous advances are happening to meet present population requirements as well as to withstand new diseases. As the population increases analyzing patients' data becomes a huge task as big data analytics helps full to manage and track those data efficiently lack of infrastructure in many areas cause providing health care a challenging task due to the constant change in the environment leads to many health crises even predicting disease based on symptoms at early stages by the doctors are less accurate because continuous mutations are also happening in disease-causing agents which makes most similar symptoms and difficult to predict. As the health sector consists of a huge amount of data, prediction of disease building a Machine learning (ML) model will give more accurate results. In this research, hybrid ml model (Naïve bayes + Ada boost) and Ensemble ml model(Random forest + Gradient boosting) and the final prediction is done by Voting classifier this model is also termed as voting Ensemble, showing huge promises and potential for accurate prediction than ML models. Performed 10 fold Cross Validation and can attain an accuracy of 85 % for Naive Bayes + Ada Boost and 92 % for Random forest + Gradient boosting,