Kernel Negative ε Dragging Linear Regression for Pattern Classification

Table 4

Accuracies (%) of different methods on the AR database.

Number of training samples per class	1	2	3	4	5
Our method	64.69	75.87	83.74	90.46	95.12
DLSR	56.83	70.01	78.34	85.89	91.37
CLSR	59.73	74.00	82.55	89.05	93.89
NNLS	59.93	67.08	78.03	85.09	92.90
-SVM	58.95	66.39	74.56	83.64	91.69
KNN	58.95	61.45	68.88	75.37	84.03
SRC	61.81	74.36	82.72	89.95	94.35
LRC	58.95	65.42	73.94	82.06	91.15
NDLR	64.00	74.87	83.38	90.02	94.63