统计分析响应面SPSSVIP免费

响应面试验设计与分析•响应面试验设计与分析:应用于按照试验设计的点进行试验,利用多项式回归方程估计出y与x1，…xp的关系.利用回归方程估计极值点、定值点，找出优化搭配。设:yx1，…，xp有关系:E(y)=f(x1，…，xp)不妨yx1，x2有关系E(y)=f(x1，x2)≈b0+b1x1+b2x2≈b0+b1x1+b2x2+b3x12+b4x22++b5x1x2以通用中心组合旋转设计为例x的标准化,1.414xxxr31.4141xxxx12011222201122111222121220122032,iiiiiijijiiiiiijijijiiiiiiijijEyfxxbbxbxbbxbxbxbxbxxbbxbxbxxbbxbxbxxbxbxx•【例】（摘自食品科学CSCD-C，2008，vol29，No.08：菊粉酶提取优化）•设为变量y与因素x1、x2、x3有关系，大致知道：下面来看他如何完成的论文.论文使用了Design-Expert软件进行了试验设计与统计分析yRyRrSSSSSSSSSSSSSS模型误差纯随机误差第五章协方差分析与SPSS应用•在方差分析中，有时试验指标y受另一个变量x的影响，又不能保证x一致，这时采用协方差分析来比较多个处理总体平均数一致否。处理观测初生重50日初生重50日初生重50日50日指标x龄重yx龄重yx龄重y龄重y观察值1.512.41.3510.21.15101.212.4xij,yij1.85121.29.41.110.619.81.3510.81.4512.21.110.41.1511.61.45101.210.31.059.21.110.61.4111.411.31.41319.21.4511.81.311.41.4513.51.4513.91.512.51.1512.81.3131.3512.81.5513.41.310.91.714.81.159.31.411.21.3511.61.412.31.19.61.511.61.158.51.4513.21.212.41.612.61.3512.21.25121.0511.21.712.51.29.31.312.81.111总和xi.,yi.13.85133.80初生重x对照配方1配方2配方31.151.15表10—2不同食欲增进剂仔猪生长情况表（单位：kg）平均1.5211.821.2810.841.3012.0718.25141.8015.40130.8015.65144.80..,iiyx初生重50日初生重50日初生重50日50日x龄重yx龄重yx龄重y龄重y1.512.41.3510.21.15101.212.41.85121.29.41.110.619.81.3510.81.4512.21.110.41.1511.61.45101.210.31.059.21.110.61.4111.411.31.41319.21.4511.81.311.41.4513.51.4513.91.512.51.1512.81.3131.3512.81.5513.41.310.91.714.81.159.31.411.21.3511.61.412.31.19.61.511.61.158.51.4513.21.212.41.612.61.3512.21.25121.0511.21.712.51.29.31.312.81.111初生重x对照配方1配方2配方3clkxy11.512.411.851211.3510.811.451011.41111.4511.811.512.511.5513.411.411.211.511.611.612.611.712.521.3510.221.29.421.4512.221.210.321.411.321.311.421.1512.821.310.921.3511.621.158.521.3512.221.29.331.151031.110.631.110.431.059.231.41331.4513.531.31331.714.831.412.331.4513.231.251231.312.841.212.4419.841.1511.641.110.6419.241.4513.941.3512.841.159.341.19.641.212.441.0511.241.111主体间效应的检验因变量:y源III型平方和df均方FSig.校正模型59.295a414.82417.013.000截距2.09212.0922.401.129x47.615147.61554.645.000clh20.43536.8127.817.000误差37.46843.871总计6410.31048校正的总计96.76347a.R方=.613（调整R方=.577）估计因变量:yclh均值标准误差95%置信区间下限上限110.339a.3359.66311.016211.074a.27110.52711.621312.149a.27011.60512.693412.312a.31211.68312.942a.模型中出现的协变量在下列值处进行评估:x=1.3156.成对比较因变量:y(I)clh(J)clh均值差值(I-J)标准误差Sig.a差分的95%置信区间a下限上限12-.735.446.107-1.634.1643-1.810*.436.000-2.688-.9314-1.973*.522.000-3.027-.92021.735.446.107-.1641.6343-1.075*.382.007-1.845-.3054-1.238*.401.004-2.048-.429311.810*.436.000.9312.68821.075*.382.007.3051.8454-.163.408.691-.986.660411.973*.522.000.9203.02721.238*.401.004.4292.0483.163.408.691-.660.986基于估算边际均值a.对多个比较的调整：最不显著差别（相当于未作调整）。*.均值差值在.05级别上较显著。【例5-1】为比较1、2、3、4四种不同肥料梨树单株平均产量，每个处理10株梨树。各株梨树的起始干周x和单株产量y列表5-1，试检验4种肥料的单株产量是否有显著差异。肥料变量x36302623263020192016y89807480856873688058x28272724252320181720y64817367776764655957x28332622232022231817y55625858665560715548x32232723272820241917y52586462545455445151表5-1梨树4种肥料比较试验干周x（㎝）单株产量y（㎏）4观测值1232.x2.y3.x3.y4x4y222()()()ijiijiyyyyyy总平方和=处理间平方和+B组间平方和+误差平方和11,1TtTtteexetSSSSSSSSSSSSSSdfnkdfkdfnkk