Frequently in econometric analysis of accounting data, one is concerned with departures from standard parametric model probability assignments. Semi- and nonparametric methods provide an alternative means to characterize data and assess parametric model robustness or logical consistency. Here, we focus on regression. That is, we examine the conditional relation between Y and X.The most flexible fit of this conditional relation is nonparametric regression where flexible fit refers to the degree of distributional or structural form restrictions imposed on the data in estimating the relationship.
KeywordsNonparametric Regression Kernel Estimator Conditional Relation Kernel Regression Generalize Cross Validation
Unable to display preview. Download preview PDF.