Employ Nature-Inspired Algorithms in Selecting the Band Width in A Polynomial Estimator

Abstract

The subject of regression analysis receives broad and explicit attention in the majority of studies, particularly those in the fields of economics and medicine. The nonparametric regression model in general and the multivariate nonparametric regression model in particular is one of the most significant regression models used in recent years which has seen significant expansion, particularly in the fields of economics and the environment. Local Polynomial Kernel estimator is one of the most used estimators in multivariate nonparametric regression model. However, this estimator is fully dependent on the matrix of the smoothing parameter, which is of crucial in the quality of the reconciliation of the estimated shape in the multivariate nonparametric regression model. In this research, employing the nature-inspired technique as Invasive Weed Optimization technique was proposed for the process of estimating the prefix parameter matrix in the estimator. The Mont-Carlo method of simulation has been used to generate data for tracking a number of multivariate regression models. Simulation results based on mean squared error have been shown by developing a benchmark that surpasses the recommended approach in comparison to alternative estimating methods.