|
Adaptive neighbor synthetic minority oversampling techniqueunder 1NN outcast handling |
|---|---|
| รหัสดีโอไอ | |
| Creator | 1. Wacharasak Siriseriwan 2. Krung Sinapiromsaran |
| Title | Adaptive neighbor synthetic minority oversampling techniqueunder 1NN outcast handling |
| Publisher | Research and Development Office, Prince of Songkla University |
| Publication Year | 2560 |
| Journal Title | Songklanakarin Journal of Science and Technology (SJST) |
| Journal Vol. | 39 |
| Journal No. | 5 |
| Page no. | 565 |
| Keyword | class imbalance problem,oversampling,SMOTE,adaptive neighbors approach,minority outcast |
| ISSN | 0125-3395 |
| Abstract | SMOTE is an effective oversampling technique for a class imbalance problem due to its simplicity and relatively highrecall value. One drawback of SMOTE is a requirement of the number of nearest neighbors as a key parameter to synthesizeinstances. This paper introduces a new adaptive algorithm called Adaptive neighbor Synthetic Minority OversamplingTechnique (ANS) to dynamically adapt the number of neighbors needed for oversampling around different minority regions.This technique also defines a minority outcast as a minority instance having no minority class neighbors. Minority outcasts areneglected by most oversampling techniques but instead, an additional outcast handling method is proposed for the performanceimprovement via a 1-nearest neighbor model. Based on our experiments in UCI and PROMISE datasets, generateddatasets from this technique have improved the accuracy performance of a classification, and the improvement can be verifiedstatistically by the Wilcoxon signed-rank test. |
