I tried to preprocess the Titanic dataset available on Kaggle. Then I fixed two predictive models (SVM and KNN) and I verified if there were significant variations in the prediction accuracy:
https://colab.research.google.com/drive/1EWL90QeF0JnYOvoqk2QpXLRtzbiQJDTW?usp=sharing
the dataset is available here:
https://drive.google.com/file/d/1q4qoplxbjh3Y4T4vxQNgmXesZWPOfMez/view?usp=sharing