Stroke Prediction
This repo contains a Machine Learning based implementation of Stroke Prediction.
● It incorporated the use of machine learning models Random Forest Classifier, SVM Classifier, XGBoost Classifier, Multilayer Perceptron Classifier and KNN Classifier.
● Incorporated SMOTE algorithm for oversampling to decrease the type - 2 Error (which is very critical from diagnosis point of view).
● Tuned hyperparameters using manual tuning and RandomizedSearchCV to decrease the Type -2 Errors and increase the accuracy.
● Finally, a Flask application was created with a User - friendly UI hosted locally.
This was a course project of the Pattern Recognition and Machine Learning course conducted by Dr. Richa Singh.