Used six different Medical and Non-Medical datasets to analyze the newly proposed regression error consistency metrics (EC) metrics with 6 ML regressor (Linear, Lasso, Ridge, RF, Knn, and SVR). All Datasets has been gone through different data preprocessing, and feature engineering steps, while all ML were tuned for every dataset. Here, we used 5-fold external holdout cross-validation and we repeated the process 50 times.
Figure: MAE VS Regression EC (House Dataset)
Figure: MAE VS Regression EC (Diabetics Dataset)
Use python3 -m run --dataset=Dataset_Name --progress to execute the main function!
You need to choose the Dataset_Name like "Boston", "Bike", etc to execute the program for the respective dataset. Additionally, if you want to see the progress bar then use --progress, otherwise ignore it.