Research Article
AWSMOTE: An SVM-Based Adaptive Weighted SMOTE for Class-Imbalance Learning
Input: majority class , minority class , nearest neighbor , , and | | Output: a new minority class | (1)A set of resamples from and are used to train SVM classifier and obtain the vector of variable weights | (2)Calculate the vector of each minority case weight : | (3)for to do | (4) Count the number of samples generated for each minority case : | (5) for to do | (6) Compute the nearest neighbors of , obtain ; | (7) for to do | (8) The element of variable of the new sample : | | Where is a random number | (9) | end for | (10) end for | (11) Add all the samples generated by to the generated minority class | (12)end for | (13) | | (14)return A new minority class |
|