THE CONSISTENCIES FOR THE KERNEL-TYPE DENSITY ESTIMATON IN THE CASE OF NA SAMPLES
NA样本概率密度函数核估计的相合性

Let be a sequence of identically distributed and negatively associated (NA) variables with probability density function f(x). Based on NA samples, the kernel estimator for f(x) is constructed, and the consistency in r-order mean, the pointwise strong consistency and strong uniform consistency are shown under suitable conditions. Finally) as the application in reliability problems, the empirical survival function and the kernel-type failure rate estimator are proposed, and the relative consistencies are disscussed.