Metaheuristic Search in Mixed Kernel and Spline Truncated Non-parametric Regression

Mustika Hadijati (University of Mataram)
Irwansyah Irwansyah (University of Mataram)
Nurul Fitriyani (University of Mataram)
Muhammad Sopian Sauri (Institut Teknologi Sepuluh Nopember)

International Journal of Computing Science and Applied Mathematics-IJCSAM
Vol. 11 No. 2 (2025)

Non-parametric regressions are widely used in data analysis because of their flexibility. Apart from their applicability, it is not easy to find the optimal parameters of the corresponding non-parametric models. This situation is caused by the nonexistence of a closed formula of the optimal parameters. In this paper, we propose a metaheuristic approach for optimal parameter search in mixed kernel and truncated spline and kernel regression. Moreover, we provide examples on how to implement the proposed algorithm to both real and simulated datasets. The results indicate that the algorithm yields highly accurate predictions for mixed truncated spline and kernel regression models.