Cost-Sensitive Support Vector Machine Using Randomized Dual Coordinate Descent Method for Big Class-Imbalanced Data Classification

Table 1

Specification of the benchmark datasets. Number of ex. is the number of example data points. Number of feat. is the number of features. Ratio is the class imbalance ratio. Target specifies the target or positive class. Number is the number of the datasets.

Dataset	Number of ex.	Number of feat.	Ratio	Positive class	Number
KDD99 (intrusion detection)	5,209,460	42	1 : 4	Normal	1
Web span	350,000	254	1 : 2	−1	2
MNIST	70,000	780	1 : 10	5	3
Covertype	581,012	54	1 : 211	Cottonwood/Willow(4)	4
SIAM1	28,596	30,438	1 : 2000	1, 6, 7, 11	5
SIAM11	28,596	30,438	1 : 716	11, 12	6