Class imbalance learning with CostSensitiveACGAN
The class imbalance problem has been recognized in many real-world applications and negatively affects machine learning performance. Generative Adversarial Networks or GANs have been known to be the next best thing in image generation. However, most GANs do not consider classes and when they do, can...
| Main Authors: | , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
Little Lion Scientific
2023
|
| Online Access: | http://psasir.upm.edu.my/id/eprint/107031/ http://psasir.upm.edu.my/id/eprint/107031/1/Class%20imbalance%20learning%20with%20CostSensitiveACGAN.pdf |
| _version_ | 1848864855418732544 |
|---|---|
| author | Abdul Halim, Athirah Hazwani Manshor, Noridayu Husin, Nor Azura |
| author_facet | Abdul Halim, Athirah Hazwani Manshor, Noridayu Husin, Nor Azura |
| author_sort | Abdul Halim, Athirah Hazwani |
| building | UPM Institutional Repository |
| collection | Online Access |
| description | The class imbalance problem has been recognized in many real-world applications and negatively affects machine learning performance. Generative Adversarial Networks or GANs have been known to be the next best thing in image generation. However, most GANs do not consider classes and when they do, cannot perform well under the imbalance problem. Based on related works, the modification of the loss function and various resampling methods have been commonly applied to counter the problem of class imbalance. In this study, CostSensitive-ACGAN is introduced which is a variation of Auxiliary Classifier GAN (ACGAN) that can work better under the class imbalance condition. This method incorporated the idea of applying costsensitive learning in the loss function to further improve the classification of minority classes. Cost-sensitive parameters are determined adaptively according to the classification error of the class to improve minority classes presence. By applying higher misclassification costs for minority classes, these instances can be magnified and recognized by the discriminator thus improving image generation altogether. This method has shown comparatively competitive results with existing benchmark models. |
| first_indexed | 2025-11-15T13:55:27Z |
| format | Article |
| id | upm-107031 |
| institution | Universiti Putra Malaysia |
| institution_category | Local University |
| language | English |
| last_indexed | 2025-11-15T13:55:27Z |
| publishDate | 2023 |
| publisher | Little Lion Scientific |
| recordtype | eprints |
| repository_type | Digital Repository |
| spelling | upm-1070312024-10-17T06:47:37Z http://psasir.upm.edu.my/id/eprint/107031/ Class imbalance learning with CostSensitiveACGAN Abdul Halim, Athirah Hazwani Manshor, Noridayu Husin, Nor Azura The class imbalance problem has been recognized in many real-world applications and negatively affects machine learning performance. Generative Adversarial Networks or GANs have been known to be the next best thing in image generation. However, most GANs do not consider classes and when they do, cannot perform well under the imbalance problem. Based on related works, the modification of the loss function and various resampling methods have been commonly applied to counter the problem of class imbalance. In this study, CostSensitive-ACGAN is introduced which is a variation of Auxiliary Classifier GAN (ACGAN) that can work better under the class imbalance condition. This method incorporated the idea of applying costsensitive learning in the loss function to further improve the classification of minority classes. Cost-sensitive parameters are determined adaptively according to the classification error of the class to improve minority classes presence. By applying higher misclassification costs for minority classes, these instances can be magnified and recognized by the discriminator thus improving image generation altogether. This method has shown comparatively competitive results with existing benchmark models. Little Lion Scientific 2023-06-30 Article PeerReviewed text en http://psasir.upm.edu.my/id/eprint/107031/1/Class%20imbalance%20learning%20with%20CostSensitiveACGAN.pdf Abdul Halim, Athirah Hazwani and Manshor, Noridayu and Husin, Nor Azura (2023) Class imbalance learning with CostSensitiveACGAN. Journal of Theoretical and Applied Information Technology, 101 (12). 5067 - 5078. ISSN 1992-8645; ESSN: 1817-3195 https://www.jatit.org/volumes/Vol101No12/19Vol101No12.pdf |
| spellingShingle | Abdul Halim, Athirah Hazwani Manshor, Noridayu Husin, Nor Azura Class imbalance learning with CostSensitiveACGAN |
| title | Class imbalance learning with CostSensitiveACGAN |
| title_full | Class imbalance learning with CostSensitiveACGAN |
| title_fullStr | Class imbalance learning with CostSensitiveACGAN |
| title_full_unstemmed | Class imbalance learning with CostSensitiveACGAN |
| title_short | Class imbalance learning with CostSensitiveACGAN |
| title_sort | class imbalance learning with costsensitiveacgan |
| url | http://psasir.upm.edu.my/id/eprint/107031/ http://psasir.upm.edu.my/id/eprint/107031/ http://psasir.upm.edu.my/id/eprint/107031/1/Class%20imbalance%20learning%20with%20CostSensitiveACGAN.pdf |