RBF Model Based on the Improved KELE Algorithm

• DOI: 10.11648/j.sjbm.20170503.12
• Publication Date: 2017-10-16T08:37:50Z
• Authors: Chen Xiu-rong
• Keywords: Kernel (algebra); Hierarchical RBF

Firstly, we use the idea of mapping by kernel function KECA to transfer original global nonlinear problem into linear one under high-dimensional feature space improve manifold learning dimension reduction algorithm LLE, then put results obtained form KELE RBF, constructing RBF model based on KELE. And choose foreign exchange rate time series verify improved model, and show that can effectively reduce samples prediction accuracy is increased obviously.