What is it about?
Many real-world datasets suffer from class imbalance, where one group is much larger than the other. This imbalance can reduce the accuracy and reliability of classification models, especially when the minority group is the most important one. In this study, we propose a new hybrid approach that combines genetic algorithms and support vector machines to improve the performance of the Synthetic Minority Over-Sampling Technique (SMOTE). By optimizing key parameters automatically, the proposed method generates more informative synthetic samples for the minority class. Our results show that the hybrid approach improves classification performance compared to standard SMOTE methods. This work can help researchers and practitioners build more accurate and fair machine-learning models when working with imbalanced data.
Featured Image
Photo by Markus Winkler on Unsplash
Why is it important?
Many decisions in health, finance, and technology rely on machine learning models trained on real-world data. When these data are imbalanced, models often perform poorly for underrepresented groups, leading to inaccurate predictions and unfair outcomes. Improving how machine learning models handle imbalanced data is therefore essential for building more reliable, fair, and trustworthy decision-support systems.
Perspectives
This work provides a practical perspective for researchers and practitioners who work with imbalanced datasets. The proposed approach can be adapted to different application areas and machine-learning models, offering a flexible way to improve classification performance. In future studies, similar optimization strategies may be extended to other data balancing techniques and real-world decision-support systems. Overall, this study highlights the value of optimization-based approaches for improving machine learning reliability in real-world applications.
Pelin Akın
Cankiri Karatekin Universitesi
Read the Original
This page is a summary of: A new hybrid approach based on genetic algorithm and support vector machine methods for hyperparameter optimization in synthetic minority over-sampling technique (SMOTE), AIMS Mathematics, January 2023, Tsinghua University Press,
DOI: 10.3934/math.2023473.
You can read the full text:
Contributors
The following have contributed to this page
