本文選題：縱向數(shù)據(jù)　切入點(diǎn)：二次推斷函數(shù)　出處：《應(yīng)用概率統(tǒng)計(jì)》2017年04期

【摘要】：針對(duì)縱向數(shù)據(jù)廣義部分線性模型,通常的做法是用樣條或核方法逼近非參部分,之后利用廣義估計(jì)方程方法(GEE)估計(jì)參數(shù)部分.本文使用B樣條逼近非參函數(shù),并基于二次推斷函數(shù)的方法對(duì)參數(shù)和非參數(shù)進(jìn)行估計(jì),并給出了估計(jì)量的大樣本性質(zhì).模擬表明本文的方法改進(jìn)了GEE的效率.
[Abstract]:For the generalized partial linear model of longitudinal data, the method of spline or kernel is usually used to approximate the non-parametric part, and then the generalized estimation equation method is used to estimate the parameter part.In this paper, we use B-spline approximation to nonparametric function, and estimate the parameter and nonparameter based on the quadratic inference function, and give the large sample property of the estimator.Simulation results show that the proposed method improves the efficiency of GEE.
【作者單位】： 北京工業(yè)大學(xué)應(yīng)用數(shù)理學(xué)院;景德鎮(zhèn)陶瓷大學(xué)信息工程學(xué)院;
【基金】：supported by the National Natural Science Foundation of China(Grant No.11571025;Key Grant No.11331011) the BCMIIS the Beijing Natural Science Foundation(Grant Nos.1142003;L140003)
【分類號(hào)】：O212.1

【相似文獻(xiàn)】

相關(guān)期刊論文 前10條

1 趙曉兵;唐玉萍;;含持久生存數(shù)據(jù)的一個(gè)縱向數(shù)據(jù)模型[J];四川文理學(xué)院學(xué)報(bào);2007年02期

2 謝婧;孫海燕;汪l，

本文編號(hào)：1705791