Rates of convergence of an adaptive kernel density estimator for finite mixture models

Kernel smoothing methods are widely used in many areas of statistics with great success. In particular, minimum distance procedures heavily depend on kernel density estimators. It has been argued that when estimating mixture parameters in finite mixture models, adaptive kernel density estimators are preferable over nonadaptive kernel density estimators. Cutler and Cordero-Braña [1996, J. Amer. Statist. Assoc. 91, 1716–1721] introduced such an adaptive kernel density estimator for the minimum Hellinger distance estimation in finite mixture models. In this paper, we investigate the convergence properties of a practical version of their adaptive estimator under some regularity conditions, and compare them with those of a nonadaptive estimator. The rates of convergence of the bias and variance of the proposed estimator are established.