Optimizing early diagnosis by integrating multiple classifiers for predicting brain stroke and critical diseases.

Machine learning has gained attention in the medical field. Continuous efforts are being made to develop robust models for early prognosis purposes. The brain is the most pivotal organ in the human body. A brain stroke is generally caused by a blockage in the brain arteries. A brain stroke is one of the primary reasons for death. Therefore, early prediction of diseases like brain stroke, heart attack can significantly help in making decisions for doctors. The research study aims to find a robust and potential technique for the early prediction of brain stroke, Alzheimer's, heart attack, cancer, Parkinson's and potentially reducing the incidence of severe post complications of the mentioned diseases. By considering the five datasets as input, machine learning models have been trained for the research study. Early prediction of brain stroke has been done using eight individual classifiers along with 56 other models which are designed by merging the pairs of individual models using soft and hard voting for brain stroke and eight individual classifiers have been used for early prediction of heart attack, cancer, Alzheimer and Parkinson's. After analyzing the results of each classifier for each disease, the proposed method, which is a pair of random forest and decision tree using a hard voting method for early brain stroke prediction, achieves the highest accuracy of 99%, which is better than all classifiers. Along with accuracy, the proposed method attained a value of 98% in precision, an outstanding 100% in recall, and 99% in F1 score. XGBoost performed best for cancer, Parkinson's, Alzeihmer's and Bernoulli naive bayes performed best in case of Heart attack .Upon comparing the values of these performance metrics, they outshine all the other model's values.