Heart-Stroke-Prediction

Here we used two datasets.

1)Dataset 1(5110 rows and 11 columns)

2)Dataset 2(31652 rows and 11 columns)

we used both benchmark and custom dataset to analyze our model result.

Oversampling and undersampling

We used 4 techniques of oversampling and undersampling to balance our dataset .

1)Borderline Smote + Random Undersampling

2)SVM Smote + Random Undersampling

3)SMOTE + Tomek

4)SMOTE + ENN

1)Pearson Correlation

2)Univariate Process

3)Extra Tree Classifier

1)KNN

2)Decision Tree

3)Support Vector Machine

4)Gaussian Naive Bayes