SMOTE: Synthetic Minority Over-sampling Technique - Implementation and experiments on datasets
Please note that SMOTE percentage should be 100% or more. The program will exit if lesser values are passed The datasets have been included in the folder itself.
To run the program for three datasets- Navigate to 'smote' folder and please run the following command.
python main.py
Datasets are already present in the folder. To select respective dataset, please have one of the following lines in main function.
generate_smote_and_compare(filename='pima-indians-diabetes.csv', smote_percentage=100) ## For Pima dataset
generate_smote_and_compare(filename='phoneme.csv', smote_percentage=200) ## For Phoneme dataset
generate_smote_and_compare(filename='covtype.csv', smote_percentage=300) ## For Forest Cover dataset