Accuracy improvement of SAR image classification method using Greedy Hierarchical Learning in comparison with Support Vector Machine

We contrast the classification of SAR images using a novel greedy hierarchical learning method with a supported vector machine. 20 dataset samples in total were taken from the Kaggle learning platform. The dataset is divided into training and testing sets. In a total of 20 datasets, 10 were used for testing and 10 for training. The classification and comparison of the recognition accuracy for support vector machine and the novel greedy hierarchical learning are done using MATLAB and synthetic aperture radar images. By fixing alpha (0.05) and power (80%), the sample size is calculated using the G power. Results: Support vector machine and novel greedy hierarchical learning both achieved classification accuracy of 80.41% and 93.08% from the MATLAB simulation, respectively. The significance found using the SPSS analysis is 0.009 (p 0.05). Conclusion: The accuracy of the novel greedy hierarchical learning outperforms the support vector machine significantly for the given data set of synthetic aperture radar image classification.