Kernel Function for Nonplanar Oscillating Surfaces in a Subsonic Flow

Objective: Lithium-ion batteries are important components used in electric automobiles (EVs), fuel cell EVs and other hybrid EVs. Therefore, it is greatly important to discover its remaining useful life (RUL). Methods: In this paper, a battery RUL prediction approach using multiple kernel extreme learning machine (MKELM) is presented. The MKELM’s kernel keeps diversified by consisting multiple kernel functions including Gaussian kernel function, Polynomial kernel function and Sigmoid kernel function, and every kernel function’s weight and parameter are optimized through differential evolution (DE) algorithm. Results : Battery capacity data measured from NASA Ames Prognostics Center are used to demonstrate the prediction procedure of the proposed approach, and the MKELM is compared with other commonly used prediction methods in terms of absolute error, relative accuracy and mean square error. Conclusion: The prediction results prove that the MKELM approach can accurately predict the battery RUL. Furthermore, a compare experiment is executed to validate that the MKELM method is better than other prediction methods in terms of prediction accuracy.

Download Full-text

Optimization of combined kernel function for SVM based on large margin learning theory

2008 IEEE International Conference on Systems, Man and Cybernetics ◽

10.1109/icsmc.2008.4811301 ◽

2008 ◽

Author(s):

Mingzhu Lu ◽

Jianbing Huo ◽

C. L. Philip Chen ◽

Xizhao Wang

Keyword(s):

Kernel Function ◽

Learning Theory ◽

Large Margin

Download Full-text

Improved Predictive Ability of KPLS Regression with Memetic Algorithms

Mathematics ◽

10.3390/math9050506 ◽

2021 ◽

Vol 9 (5) ◽

pp. 506

Author(s):

Jorge Daniel Mello-Román ◽

Adolfo Hernández ◽

Julio César Mello-Román

Keyword(s):

Kernel Function ◽

Linear Method ◽

Predictive Ability ◽

Feature Space ◽

Memetic Algorithms ◽

Least Squares Regression ◽

Optimal Parameters ◽

Number Of Components ◽

Kernel Partial Least Squares ◽

Dependent Variables

Kernel partial least squares regression (KPLS) is a non-linear method for predicting one or more dependent variables from a set of predictors, which transforms the original datasets into a feature space where it is possible to generate a linear model and extract orthogonal factors also called components. A difficulty in implementing KPLS regression is determining the number of components and the kernel function parameters that maximize its performance. In this work, a method is proposed to improve the predictive ability of the KPLS regression by means of memetic algorithms. A metaheuristic tuning procedure is carried out to select the number of components and the kernel function parameters that maximize the cumulative predictive squared correlation coefficient, an overall indicator of the predictive ability of KPLS. The proposed methodology led to estimate optimal parameters of the KPLS regression for the improvement of its predictive ability.

Download Full-text