This paper is concerned with nonparametric binomial regression. Two kernel-based binomial regression estimators and their bias-adjusted versions are proposed, whose kernels are weighted by the inverses of variance estimators of the observed proportion at each covariate. Asymptotic theories for deriving asymptotic mean squared errors (AMSEs) of proposed estimators are developed. Comparisons with other estimators discussed by several authors are implemented through the AMSEs. From these considerations, together with the simulation results, the advantages of our weighting scheme are reported.

This paper presents a kernel smoothing method for multinomial regression. A class of estimators of the regression functions is constructed by minimizing a localized power-divergence measure. These estimators include the bandwidth and a single parameter originating in the power-divergence measure as smoothing parameters. An asymptotic theory for the estimators is developed and the bias-adjusted estimators are obtained. A data-based algorithm for selecting the smoothing parameters is also proposed. Simulation results reveal that the proposed algorithm works efficiently.

In nonparametric binomial regression, the weighted kernel estimator of the regression function and its efficient bias-adjusted version have been proposed by Okumura and Naito (2004) with consideration to differences of variances of observed response proportions at covariates. The aim of this article is to propose an effective data-based method for bandwidth selection of the bias-adjusted estimator. The proposed method is developed through three steps: the plug-in method by Ruppert et al. (1995), a scale adjustment suggested by Yang and Tschernig (1999), and an effective use of the approach discussed by Grizzel et al. (1969) for the rule-of-thumb part. Theoretical considerations on the asymptotic performance of the selected bandwidth are given under the situation where the numbers of covariates and responses observed at each covariate increase.

In this paper, we consider kernel-based estimators in the nonparametric binary regression problem with multidimensional covariates. We propose a local linear type estimator of the response probability function with kernel weighted at each observed covariate. In addition, we discuss the rule of thumb bandwidth selector and the plug-in bandwidth selector. The efficiency of the weighted local linear estimator is determined from results of asymptotic properties and our simulation study.

This article considers identifying the existence of the local likelihood estimator in binary regression. A simple method for the identification is proposed, which is derived by recognizing the problem as a linear complementarity problem through a support vector machine problem with soft margins.