“…Our goal is to estimate the unknown vector function F ( X ). To this end, similarly to Yu, Wang & Shi (), Farnoosh & Mortazavi () and Farnoosh, Hajebi & Samadi () for the univariate case, we assume a parametric framework for the vector autoregression function F ( X ), that is, G ( X , Θ ). However, unlike in Yu, Wang & Shi () and Farnoosh & Mortazavi (), instead of making a crude guess of the true density function F (·), we use a multivariate Taylor series expansion approximation to estimate the parametric function G ( X , Θ ).…”