International Journal of Applied Mathematics and Machine LearningAuthor's: Lichao Xia, Kaihui Li, Wenjing Xia and Jianqiang Gao
Pages: [87] - [101]
Received Date: June 15, 2017; Revised July 03, 2017
Submitted by:
DOI: http://dx.doi.org/10.18642/ijamml_7100121844
In this paper, an efficient, novel and simple method is proposed to approximately solve linear ordinary differential equations (ODEs), emphasis is put on high-order linear ODEs with variable coefficients. What’s more, by using the least squares support vector regression (LS-SVR) method, numerical solution based on LS-SVR for high order linear ODEs with variable coefficients can be easily solve. The following illustrative examples are given to demonstrate the effectiveness and high accuracy of we proposed method.
SVR, variable coefficients, numerical solution, LS-SVR, ODE.
