资源描述:
《多元正态分布的假设检验法分析ppt课件.ppt》由会员上传分享,免费在线阅读,更多相关内容在教育资源-天天文库。
1、多元正态分布的假设检验§4.1单个总体均值向量的推断prociml;n=20;p=3;x={3.748.59.3,5.765.18.0,3.847.210.9,3.253.212.0,3.155.59.7,4.636.17.9,2.424.814.0,7.233.17.6,6.747.48.5,5.454.111.3,3.936.912.7,4.558.812.3,3.527.89.8,4.540.28.4,1.513.510.1,8.556.47.1,4.571.68.2,6.552.810.9,4.144.111.2,5.54
2、0.99.4};m0={45010};ln={[20]1};x0=(ln*x)/n;printx0;xm=x0-m0;printxm;mm=i(20)-j(20,20,1)/n;a=x`*mm*x;printa;ai=inv(a);printai;dd=xm*ai*xm`;d2=(n-1)*dd;t2=n*d2;f=(n-p)*t2/((n-1)*p);printddd2t2f;p0=1-probf(f,p,n-p);printp0;fa=finv(0.95,p,n-p);beta=probf(fa,p,n-p,t2);print
3、fabeta;quit;TheSASSystem08:48Wednesday,March10,20084X04.6445.49.965XM0.64-4.6-0.035A54.708190.19-34.372190.193795.98-107.16-34.372-107.1668.9255AI0.0308503-0.0011620.0135773-0.0011620.0003193-0.0000830.0135773-0.0000830.0211498DDD2T2F0.02562830.48693869.73877292.90454
4、63P00.0649283FABETA3.19677680.3616381二单个总体均值分量间结构关系的检验是取自该总体的样本。检验:1、问题引入例设与上面的假设等价的是,寻找常数矩阵注:矩阵C不是唯一的,在例4.2.1中,假定人类的体形有这样一个一般规律的身高、胸围和上臂围平均尺寸比例为6:4:1。检验比例是否符合这一规律。检验:则上面的假设可以表达为2、统计量及方法其中C为一已知的k×p阶矩阵,k
5、算成F统计量。对给定的显著性水平α,检验的规则某地区农村男婴的体格测量数据如下编号身高(cm)胸围(cm)上半臂长(cm)17860.616.527658.112.539263.214.548159.014.058160.815.568459.514.0检验三个指标的均值是否有关系prociml;s={31.6008.0400.500,8.0403.1721.310,0.5001.3101.900};mu={82.0060.2014.50};c={2-30,10-6};a=c*t(mu);d=c*s*t(c);g=inv(d);T
6、=6#(t(a)*g*a);f=((6-2)/(2*(6-1)))*T;PrintT,f;p0=1-probf(f,2,6-2);printp0;fa=finv(0.95,2,6-2);printfa;Quit;T=47.143TheSASSystem08:48Wednesday,March10,200818T47.143404F18.857362P00.0091948FA6.9442719§4.2两个总体均值的检验一、两个独立样本的情形与一元随机变量的情形相同,常常我们需要检验两个总体的均值是否相等。设从总体,中各自独立地抽取样
7、本和,。考虑假设根据两个样本可得μ1和μ2的无偏估计量为其中当原假设为真的条件下,检验的规则为:datad331;inputtypex1-x4;cards;165352560175502055160453565175404070170303050155403565160453060165402560160503070155553575255554065250604570245453575250505070255503075260404560265554575250603580240453065245504570;prociml;n
8、=10;m=10;p=4;used331(obs=10);xx={x1x2x3x4};readallvarxxintox;printx;ln={[10]1};x0=(ln*x)/n;printx0;mx=i(n)-j(n,n,1)/n;a1=x`*