-混合样本下线性模型误差分布的相合核估计

On the Convergent Kernel Estimation of Error Distribution in Linear Model Under -Mixing Samples

对于有公共的未知分布密度f(x)的一个线性模型的误差序列,首先利用收敛系统原理证明了回归系数β的最小二乘估计的相合性,然后由-混合序列的矩不等式,得到了f(x)的核估计■n(x)的逐点弱相合性。

For a linear model,the error sequence has a common unknown density f（x）.This paper first gets the convergence of β′s L.S.estimation using the conception of convergent system,and then proves the convergence of f（x）′s kernel estimation fn（x） under ρ-mixing sequence,in probability with the help of moment inequality.