测线调查样本密度推定中核函数选择的几点注记

SOME NOTES ON DENSITY ESTIMATION TO KERNEL FUNCTION FOR LINE TRANSECT SAMPLING

对于半正态探知函数,使用两个最佳带幅宽和六个属于相同的函数族、不同的核函数 (其中 3个是经常被使用、著名的核函数,另外 3个是我们作成的核函数)来进行密度推定,通过计算机数值模拟产生测线调查样本;用统计量RRMSE和RB对密度推定值进行评价,来验证核函数的优劣。

Compare and consider kernel estimation that two optimal bandwidths could be chosen and six kernel functions (among them three kernels are often used and famous kernel, other three kernel functions are done by us), which belong to the same function family was used, for half-normal detection function. Through the simulations and prove that about different conditions, what kernel function is more better.