STRONG UNIFORM CONSISTENCY RATES FOR CONDITIONAL FUNCTION ESTIMATORS WITH ■-MIXING SAMPLE

摘要

Strong uniform consistency rates are given for kernel type estimatorsof the conditional function with （?）-mixing sample.Especially,for nonparametricestimators of kernel density,the regression function when Y is bounded,conditionaldf’s,L-smoothing and M-smoothing,we obtain the same rate O(（n/log n）-1/3)as in the i.i.d.sample established by H（?）rdle,Janssen and Serfling.