Detect outlier first, if you data set maybe contain outlier.
Try the isolationForest method, it's fast and efficient to detect the outliers.
wolfe
- 637
- 1
- 6
- 12