Nishika, Nandita and Sharma, Anuraganand (2024) Exploring MCMC Guided GAN and Comparative Analysis for Uneven Class Distribution. [Conference Proceedings]
Full text not available from this repository. (Request a copy)Abstract
Uneven class distribution in data sets poses a significant challenge in the classification model which causes the sub-optimal performance and a biased model towards the majority class. This paper investigates the sampling method of Generative Adversarial Network (GAN) on a Bayesian Framework utilizing Markov Chain Monte Carlo (MCMC) to enhance the sampling. MCMC is employed to explore the solution space effectively, while GAN generates synthetic samples to balance the class distribution. The integration of these two techniques aims to enhance the model's robustness and generalization capabilities. Experimental results demonstrate potential improvements in classification accuracy, highlighting the potential of the proposed method in handling uneven class distribution that is MCMC Guided GAN. This method along with other methods including Synthetic Minority Over-sampling Technique (SMOTE), Smotified-GAN, GAN and MCMC. The promising results obtained highlight the potential of the MCMC Guided GAN method as a valuable tool in addressing the challenges associated with imbalanced datasets. This research contributes to the advancement of techniques for more effective and equitable classification models.
| Item Type: | Conference Proceedings |
|---|---|
| Additional Information: | IEEE Xplore page https://ieeexplore.ieee.org/xpl/conhome/10729761/proceeding |
| Subjects: | Q Science > QA Mathematics > QA75 Electronic computers. Computer science |
| Divisions: | School of Information Technology, Engineering, Mathematics and Physics (STEMP) |
| Depositing User: | Anuraganand Sharma |
| Date Deposited: | 30 Jan 2025 02:23 |
| Last Modified: | 30 Jan 2025 02:23 |
| URI: | https://repository.usp.ac.fj/id/eprint/14792 |
Actions (login required)
 |
View Item |