《多元统计试验四多元方差分析.docx》由会员分享，可在线阅读，更多相关《多元统计试验四多元方差分析.docx（9页珍藏版）》请在三一文库上搜索。
1、桂林电子科技大学数学与计算科学学院实验报告院系数学与计算科学学院学号姓名成绩课程名 称应用多元统计实验实验项目名 称实验四多兀方差分析一，实验目的用SAS软件进行多个总体均值的推断【，实验原理anova过程功能：主要用于处理均衡设计，分析各总体均值之间是否有差异格式：PROC ANOVA 选择项 1CLASS 变M ;MODEL因变M二效应变M ;MANOVA H=效应变量;RUN ;三，实验内容地质勘测中，在 A,B,C三个地区采集了一些岩石，测量其中化学成分，其数据见下表SiO2FeOK2OA地区47.225.060.1047.454.350.1547.526.850.1247.864.1
2、90.1747.317.570.18B地区54.336.220.1256.173.310.1554.402.430.2252.625.920.12C地区43.1210.330.0542.059.670.0842.509.620.0240.779.680.04假定三个地区的岩石成分遵从N3(O(i =1,2,3)(- 0.05)检验 HO : i 仁1 2= 1 3,H1 : 11, 1 2, 1 3中至少有两个不相等四，实验过程原始记录(数据，图表，计算等)data shiya n4;in put g x1-x3;|cards ;147.225.060.1147.454.350.15147.5
3、26.850.12147.864.190.17147.317.570.18254.336.220.12256.173.310.15254.42.430.22252.625.920.12343.1210.330.05342.059.670.08342.59.620.02340.779.680.045proc anova data =shiya n4;class g;model x1-x3=g;manova h =g;run ;(1)MANOVA Test Cr iter la andF Approximations for the Hypothesis ofNo Overal I W Ef f
4、ectH 二 Anova SSCP Matrix for £Statistict iIks' LambddPi 1 lai TraceHotel 1 ing-Lawley TraceRoy's Greetest RootS=2M=0昨3ValueF Value Num DFDen DFPr >FO.OW09794IS.39S16<.00011.440347657J28180.000332.8955232142.17S9.0309<*000132.10643753嗡聪99<,0001E = Error SSCP MatrixNOTE : F St
5、at I st ic for Roy I Greatest Root is an upper bound.NOTE F Statistic for Wilks' Lambda is exa?t三个地区的岩石成分 有 十分显著的差由于各种检验的p值都小于0.05,所以拒绝原假设，认为 异。(2)为了解这三种销售方式的显著差异究竟是由哪些商品引起的，我们对这四种商品分别用一元方差分析方法进行检验分析，即查看由model x1-x3=g;所输出的结果1.ependent Variable : xlSourceModelErrorCorrected TotalSum of Square;,DF
6、302.944350029.6180800w312.4634808M?n Square F V?lue151.47217540.9519080159.12Pr > F<.OQQ112R-SqoareCoeffVarRoot MSE0.969535金田0.175658xl Mean47.94?C3SourceDFAnova SSI*e4n Squtra F ValuePr > F302.3443503151.4721754159.12coooi2.leperidenl V&rtableSourcsModelErrorCorrected IctalDF103.Wg ofS
7、quaresS4.682487C920.29982C0012M.98230769Me*n Squira92.841248862.02998200Sourceg>ependeri t YariablE :SciurceModelErrorCor recTotalF VAlde Pr > F15,330.000BRASquare Coeff Var Root NSE x2 Me*nO.7UI2321.738521.424774 畀血 38 时DFAnovu SS64.68248769MeAr Square32.34124385F ?aluePr > F15.93O.QOOBOF4
8、？ af SquaresMean SquareF Yftlue2D.Q2800C82D.014003U1CJ1100.013070000.0013070()120.04K7891Pr > F0.钳 1$1&白巾”和册& 0.0361520.I1S323SourceOFAnova SSMean Square F Value Pr > F520.02A00892Q.0140A4?10."0.00AR-Squire Coeff Var Root MSE x3 Mean66 63 156 56 162 64 166 68 165 66 16160160 64 16
9、0 57 158 60 158 59 160 57 253 55 266 65 260 53 259 58 258 54 260 56 254 59 262 66 259 61 265 70 3上面的输出结果表明：岩石成分A有显著差异（p=0.001）,岩石成分B有十分显著的差异（p=0.0008）岩石成分C有显著差异（p=0.0033）。即三个地区的岩石有差异是由ABC三种岩石成分引起的课后练习程序data shiya n4;in put n x1-x2 g ;cards ;15959111260651123586211345959114550481156596511676262117863
10、621189687211910637212017069211269682123656521346260214559562156555821676058217858642189656721910676222016363311256 573362 623459 583562 583650 573763 633861 623955 5931063 633126464313656531467673155555316565631765673186265319556132058583proc anova data =shiya n4;class g;model x1-x2=g;man ova h =g;/*要求做多元方差分析 */run ;quit ;五，实验结果分析或总结通过这次实验，我学会了如何进行多元方差分析。