Int. J. of Global Warming   »   2016 Vol.9, No.2

 

 

Title: Utilising key climate element variability for the prediction of future climate change using a support vector machine model

 

Authors: Adamu Abubakar; Haruna Chiroma; Akram Zeki; Mueen Uddin

 

Addresses:
Department of Information Systems, International Islamic University Malaysia, 50728 Gombak, Kuala Lumpur, Malaysia
Department of Artificial Intelligence, University of Malaya, 50603 Pantai Valley, Kuala Lumpur, Malaysia; Federal College of Education (Technical), Gombe, Nigeria
Department of Information Systems, International Islamic University Malaysia, 50728 Gombak, Kuala Lumpur, Malaysia
Department of Information Systems, International Islamic University Malaysia, 50728 Gombak, Kuala Lumpur, Malaysia

 

Abstract: This paper proposes a support vector machine (SVM) model to advance the prediction accuracy of global land-ocean temperature (GLOT), which is globally significant for understanding the future pattern of climate change. The GLOT dataset was collected from NASA's GLOT index (C) (anomaly with base: 1951-1980) for the period 1880 to 2013. We categorise the dataset by decades to describe the behaviour of the GLOT within those decades. The dataset was used to build an SVM Model to predict future values of the GLOT. The performance of the model was compared with a multilayer perceptron neural network (MLPNN) and validated statistically. The SVM was found to perform significantly better than the MLPNN in terms of mean square error and root mean square error, although computational times for the two models are statistically equal. The SVM model was used to project the GLOT from the pre-existing NASA's GLOT index (C) (anomaly with base: 1951-1980) for the next 20 years (2013-2033). The projection results of our study can be of value to policy makers, such as the intergovernmental organisations related to environmental studies, e.g., the intergovernmental panel on climate change (IPCC).

 

Keywords: global land-ocean temperature; GLOT; climate change indicators; support vector machines; SVM; key climate element variability; climate change prediction; multilayer perceptron neural networks; MLPNNs; mean square error; root MSE.

 

DOI: 10.1504/IJGW.2016.074952

 

Int. J. of Global Warming, 2016 Vol.9, No.2, pp.129 - 151

 

Available online: 26 Feb 2016

 

 

Editors Full text accessAccess for SubscribersPurchase this articleComment on this article