Imbalanced Learning in Land Cover Classification: Improving Minority Classes’ Prediction Accuracy Using the Geometric SMOTE Algorithm
In spite of its importance in sustainable resource management, the automatic production of Land Use/Land Cover maps continues to be a challenging problem. The ability to build robust automatic classifiers able to produce accurate maps can have a significant impact in the way we manage and optimize natural resources. The difficulty in achieving these results comes from many different factors, such as data quality and uncertainty, among others. In this paper, we address the imbalanced learning problem, a common and difficult problem in remote sensing that affects the quality of classifiers. Having very asymmetric distributions of the different classes constitutes a significant hurdle for any classifier. In this work, we propose Geometric-SMOTE as a means of addressing the imbalanced learning problem in remote sensing. Geometric-SMOTE is a sophisticated oversampling algorithm which increases the quality of the generated instances over previous methods, such as Synthetic Minority Oversampling TEchnique. The performance of Geometric-SMOTE, in the the LUCAS dataset, is compared to other oversamplers using a variety of classifiers. The results show that Geometric-SMOTE significantly outperforms all the other oversamplers and improves the robustness of the classifiers. These results indicate that, when using imbalanced datasets, remote sensing researchers should consider the use of these new generation oversamplers to increase the quality of the classification results.