Description:
This notebook provides a comprehensive preprocessing pipeline for a stroke prediction dataset. Key preprocessing steps include handling missing data, encoding categorical variables, and applying undersampling techniques to balance the class distribution. The final output is a dataset ready for subsequent machine learning model training.
Description:
This notebook is designed to enhance a stroke prediction dataset by introducing controlled noise to the 'Age' column. The noise addition is executed iteratively across 10 runs, with noise levels incrementally increased by 2% in each iteration.
Description:
In this notebook, a RandomForest classifier is implemented to predict stroke risk based on patient attributes. Additionally, the notebook introduces a J48 decision tree model to compare against the RandomForest model. The notebook also includes visualizations and performance metrics for model evaluation, making it a complete solution for stroke risk prediction analysis.