A Novel SMOTE-Based Classification Approach to Online Data Imbalance Problem

Table 4

Comparative results on Macao forecasting data in 2010.

Algorithms	DGSMOTE	OS-ELM	MCOS-ELM	LS-SVM
Testing time (s)	0.0399	0.0468	0.0468	1.6224
Minority training accuracy	0.9655	0.5775	0.7037	0.7009
Majority training accuracy	0.9481	0.9771	0.9746	0.9286
Whole testing accuracy	0.9546	0.9432	0.9368	0.9025
-mean	0.9472	0.7403	0.8213	0.7655