Ensemble Learning Technique for Cloud Classification
Aarti Kumthekar1, Ramachandra Reddy G2
1Aarti Kumthekar, Department of Communication, School of Electronics Engineering (SENSE), Vellore Institute of Technology, Vellore, India.
2Dr. Ramachandra Reddy G*., Senior Professor, Department of Communication, School of Electronics Engineering (SENSE), Vellore Institute of Technology, Vellore, India.
Manuscript received on November 22, 2019. | Revised Manuscript received on December 15, 2019. | Manuscript published on December 30, 2019. | PP: 2582-2587 | Volume-9 Issue-2, December, 2019. | Retrieval Number: B3957129219/2019©BEIESP | DOI: 10.35940/ijeat.B3957.129219
Open Access | Ethics and Policies | Cite | Mendeley
© The Authors. Blue Eyes Intelligence Engineering and Sciences Publication (BEIESP). This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/)
Abstract: Automatic cloud classification is one of the important areas of remote sensing for metrological applications. Machine learning and deep learning techniques have been used for automatic classification of the cloud type. Several pretrained models are developed using convolutional neural network (CNN), which is part of deep learning. The classification performance of pretrained networks can be further improved using ensemble methods. Ensemble learning can perform better than single learner. In this paper, we proposed two different ensemble learning techniques: ensemble of CNN and ensemble of classifier. In first approach, CNN ensemble is performed, where the features extracted by two or more CNN are combined together using single classifier. The second method is to ensemble the predictions of different classifiers produced by a single or multiple CNN. The accuracy of cloud classification of the proposed methods has improved compared to without ensemble of pretrained networks.
Keywords: Pretrained network, Cloud classification, Ensemble learning