Support Vector Machine with Restarting Genetic Algorithm for Classifying Imbalanced Data

Abstract

Algorithms for data classification are normally at their high performance when the dataset has good balance in which the number of data instances in each class is approximately equal. But when the dataset is imbalanced, the classification model tends to bias toward the majority class. The goal of imbalanced data classification is how to improve the performance of a model to better recognize data from minority class, especially when minority is more interesting than the majority data. In this research, we propose technique for balancing data with hybrid resampling techniques and then perform parameter optimization with restarting genetic algorithm. The optimized parameters are for support vector machine to induce efficient model for recognizing data in minority class, whereas maintaining overall accuracy. The experimental results show that the proposed technique has high performance than others.

Deep Autoencoder Networks Optimized with Genetic Algorithms for Efficient ECG Clustering

“…However, the design of these network models are limited by the fact that an inappropriate layout of the network may yield suboptimal results. Therefore, some researchers employ genetic algorithms (GAs) to find the best network structure [22]- [25].…”