数理统计程序总结.doc

数理统计程序总结 （每个题中的数据来自书本的习题或例题，考试时请根据具体的题而改变数据,请大家先运行一下，分析运行结果，如有疑问，群上交流。） 题型一：求做频率直方图(p132) data ex; input x@@; cards; 70 72 94 24 57 68 40 99 34 26 48 105 45 55 73 58 101 94 62 84 61 40 77 66 98 50 58 76 81 54 75 118 69 25 97 76 100 43 60 34 ; proc gchart; vbar x; run; 题型二：给一组数据，求其数字特征(p137-138) 不需要频数 data ex; input x@@; cards; 1 2 2 3 3 3 4 5 6 7 8 ; proc univariate vardef = n; var x; run; proc means data=ex; var x; output; run; 需要频数 data ex; input x f @@; cards; 5.5 4 7.5 11 9.5 17 11.5 23 13.5 18 15.5 14 17.5 10 19.5 3 ; proc univariate vardef = n; var x;freq f; run; proc means data=ex; var x; freq f; output; run; 题型三：已知样本的总数，平均数，样本标准差，求其总体均值和总体方差的0.95置信区间。（p170） 求其总体均值的0.95置信区间。 data val1; xbar=1.57;s=0.077;n=25; u=tinv(0.975,24); delta=u*s/sqrt(n-1); c1=xbar-delta; c2=xbar+delta; run; proc print data=val1; var c1 xbar c2; run; 求其总体方差的0.95置信区间。 data val2; xbar=1.57;s=0.077;n=25; u1=cinv(0.975,24);u2=cinv(0.025,24); delta=n*s*s; c1=delta/u1; c2=delta/u2; run; proc print data=val2; var c1 xbar c2; run; 题型四：已知样本的各个数据，求Ex和Dx的双侧0.95置信区间。(p170-2) data ex; input x@@; cards; 24.3 20.8 23.7 19.3 17.4 ; proc capability data=ex noprint; intervals x/method=4,6; run; 题型五：已知两个样本的各个数据，求他们的均值差的0.95置信区间。(p170-5) data ex1; input weight@@; cards; 210 235 239 241 244 256 ; proc means data=ex1; var weight; output out=tval1 mean=xbar css=ss1 n=n; Run; data ex2; input weight@@; cards; 203 338 358 271 ; proc means data=ex2; var weight; output out=tval2 mean=ybar css=ss2 n=m; Run; data tval3; set tval1; set tval2; df=n+m-2;t=tinv(0.975,df); lt=xbar-ybar-t*sqrt((ss1+ss2)/df)*sqrt(1/n+1/m);ut=xbar-ybar+t*sqrt((ss1+ss2)/df)*sqrt(1/n+1/m); Run; proc print data=tval3;var lt ut; run; 题型六：一个正态总体均值作假设检验。(p186-2) data ex; input x @@; y=x-300; cards; 308 305 311 298 315 300 321 294 320 ; proc means data=ex mean std t prt;var y;run; 题型七：两个正态总体均值作假设检验。(p187-9) data ex;do a=1 to 2; do i=1 to 10; input x @@; output;end;end; cards; 101 100 99 99 98 100 98 99 99 99 100 98 100 99 98 99 98 98 99