余弦模型的非线性回归拟合法(NLR)计算传染病高峰月日的探讨

A NLR—METHOD OF FITTING CoSINE MODEL AND ITS PEAK—DAY ESTIMATE FOR IHE INCIDENCE OF INFECTIOUS DISEASES

摘要: 本文对《疾病监测》1996；(6)：228的数据，进行两种方法(圆分布法与余弦模型法)的再分析：重点介绍以SPSS/PC+软件包的NLR命令拟合余弦模型的简便方法；说明两法等价的原理与条件，并与原作者商榷。

Abstract: The datafromDISEASE SURVEILLANCE1996；11(6)：228were reanalyzed by the authors in this paper The cosine(COS)model and circular distribution(CD)model were used for fitting the equations and estimating the peakday．The former method could abso obtained the expected values.Because the two models were not so easy to be fitted．a simple procedure of NLRmethod from SPSS/PC+4.0 package Was written by the authors to calculate the needed parameters，the peakday and the expected values of COSmode；A better equation of COSmodel which was equivalent to cDmodel WS suggested