An Efficient Algorithm for the Detection of Outliers in Mislabeled Omics Data

Table 6

Results of Ensemble three models for the original TNBC data and subset with outliers removed.

Dataset	EN		SPLS-DA		SGPLS
	Model size	MR^#	Model size	MR	Model size	MR
Original data	175	0.012	22	0.064	33	0.059
Subset	83	0.000	87	0.008	16	0.015
Subset^##	49	0.001	38	0.014	55	0.013

This subset is the original dataset after removing 68 outliers identified by enetLTS. ##This subset is the original dataset after removing 47 outliers identified by MTL-EN. Model size: number of variables; #MR: misclassification rate.