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

Figure 5

The distribution of the forecasting data (a) before and (b) after the first granulation division and (c) the result after using SMOTE.

(a)

(b)

(c)