Partial Linear Models With Fuzzy Data

In this paper, we introduce a semiparametric ridge regression approach under imprecise circumstances. To end this purpose, A fuzzy regression model is used in evaluating the functional relationship between the dependent and independent variables in a partial linear model. Most fuzzy regression models are considered to be fuzzy outputs but crisp inputs. Thus we employ fuzzy least squares methods for the analysis for the partial linear model when the outputs are assumed to be fuzzy. To accommodate the proposed methodology for facing with sophisticated environments, some additional artificial linear restrictions are imposed to the whole parameter space. Semiparametric ridge and non-ridge fuzzy type estimators, in a restricted manifold are then defined. Asymptotic distributional bias and risk are also derived and some comparison results are given. A Monte Carlo simulation study is also conducted to estimate the parametric and non-parametric parts in the fuzzy regression model.

کلیدواژه ها:

Asymptotic distributional risk ، Linear restrictions ، Kernel smoothing ، fuzzy regression estimator ، Semiparametric.