Abstract : We propose a hybrid of haar wavelet decomposition, relevance vector machine, and adaptive linear neural network (HWD-RVMALNN) for the estimation of climate change behavior. The HWD-RVMALNN is able to improve estimation accuracy of climate change more than the approaches already discussed in the literature. Comparative simulation results show that the HWD-RVMALNN outperforms cyclical weight/bias rule, Levenberg-Marquardt, resilient back-propagation, support vector machine, and learning vector quantization neural networks in both estimation accuracy and computational efficiency. The model proposes in this study can provide future knowledge of climate change behavior. The future climate change behavior can be used by policy makers in formulating policies that can drastically reduce the negative impact of climate change, and be alert on possible consequences expected to occur in the future.
https://hal.inria.fr/hal-01397202 Contributor : Hal IfipConnect in order to contact the contributor Submitted on : Tuesday, November 15, 2016 - 3:37:28 PM Last modification on : Wednesday, November 16, 2016 - 1:04:11 AM Long-term archiving on: : Thursday, March 16, 2017 - 6:51:57 PM
Haruna Chiroma, Sameem Abdulkareem, Adamu I. Abubakar, Eka Novita Sari, Tutut Herawan, et al.. Hybridization of Haar Wavelet Decomposition and Computational Intelligent Algorithms for the Estimation of Climate Change Behavior. 2nd Information and Communication Technology - EurAsia Conference (ICT-EurAsia), Apr 2014, Bali, Indonesia. pp.238-247, ⟨10.1007/978-3-642-55032-4_23⟩. ⟨hal-01397202⟩