《统计学课程设计——回归分析.docx》由会员分享，可在线阅读，更多相关《统计学课程设计——回归分析.docx（15页珍藏版）》请在三一办公上搜索。
1、课程名称：统计学设计题目：相关分析院系：土木工程系专业：工程造价年级：2010级工程造价1班姓名：古雨卉指导教师：刁明碧西南交通大学峨眉校区2012年6月课程设计任务书专业-工程造价姓名.占雨卉学号2010口54开题日期：2012年6月完成日期：2012年6月题目相关分析一、设计的目的相关分析二、设计的内容及要求1 .收集12组以上经济/技术数据；2 .进行相关分析；3 .用计算机设置步骤；三、指导教师评语四、成绩指导教师（签章）年月日3-11支出法国内生产总值本表按当年价格计算。年份支出法国内生奥本形成率%)最终消费率(%)产总值(iS)最终消费资本形成总额货物和服务净出口19783605.
2、62239.11377.9-11.438.061.819794073.92619.41474.2-19.736.564.919804551.32976.11590.0-14.835.265.919814901.43309.11581.011.332.568.119825489.23637.91760.291.133.268.719836076.34020.52005.050.833.867.719847164.44694.52468.61.334.465.519858792.15773.03386.0-366.937.864.4198610132.86542.03846.0255.237.76
3、4.1198711784.07451.24322.010.836.162.3198814704.09360.15495.0-151.136.862.7198916466.010556.56095.0-185.536.062.4199018319.511365.26444.0510.334.761.3199121280.413145.97517.0617.534.860.8199225863.615952.19636.0275.536.259.9199334500.620182.114998.0-679.543.358.3199447110.927216.219260.6634.141.258.
4、2199559404.934529.423877.0998.540.859.0199669366.041039.526867.21459.339.660.5199776077.24476S.228564.02745.038.259.9注：支出法国内生产总值不等于国内生产总值是由于计算误差的影响。S电凄公据正文国内生产总值(2011年11季度)绝对语(亿元)比去年同财长3国内生户物|7156375T第一产处411120455BZtik220591.6106MZFltIa02Ql8.9注1：绝封曲将限价计r用长速度挤不JUn计篇。注2：谈襄为减版题黑。CDPHtmeei：环H长变度为好季节期与上一
5、冬度?H匕的博代建3bSSSKftfci11素CF矽喝除权所毒：中尊人R4M(三国T典计0我国的国内生产总值与最终消费额单位：亿元年份GDP最终消费额19783605.62239.119794092.62633.719804592.93007.919815008.83361.5198255903714.819836216.24126.419847362.74846.319859076.75986.3198610508.56821.8198712277.47804.6198815388.69839.5198917311.311164.2199019347.812090.5199122577.41
6、4091.9199227565.217203.3199336938.121899.9199450217.429242.2199563216.936748.2199674163.643919.5199781658.548140.6199886531.651588.219999112555636.920009874961516200110902866933.92002120475.671816.52003136634.877685.52004160800.187552.62005187131.299051.32006222240112631.92007265833.9131510.12008314
7、901.3152346.62009346316.6166820.12010394307.6186905.3注：资本形成率指资本形成总额占支出法国内生产总值的比重；最终消费率指最终消费支出占支出法国内生产总值的比重。表2-1用excel进行回归分析的步骤:Stepl:选择数据工具排序的选全部成新插入页面布局数据分圻Step2:在分析工具中选择分析工具面Step3:选择最终消费额为X值，GDP为Y值，则应在对话框中作如下选择:回归二输入Y值输入区域Il）：$C$4:$C$36匪JX值输入区域&）：；标志Qi）2置信度电）取消SB$4:$B$36常数为零1)95%国I帮助Qi）输出选项Q输出区域）：
8、$雄41|新工作表组任）：新工作薄置）残差团残差残差图标准残差复）也线性拟合图建）正态分布J正态概率图国）Step4:单击确定得到下面的结果。回归统计MultipleR0.995955304RSquare0.991926967AdjustedRSquare0.991666546标准误差4788.063985观测值33方差分析dfSSMSFSignificanceF回归分析18.73E+108.73E+103808.9455.17E-34残差317.11E+O822925557总计328.8E+10Coefficients标准误差tStatP-valueLower95%Upper95%下限95.
9、0%上限95.0%Intercept4741.3074941097.584.3197820.0001492502.7786979.8372502.7786979.837XVariable10.4830670750.00782761.716655.17E-340.4671030.4990310.4671030.499031RESIDUALOUTPUTPROBABILITYOUTPUT观测值预测Y残差标准残差百分比排位Y16483.054139-4243.95-0.900541.5151522239.126718.307804-4084.61-0.866734.5454552633.736959.
10、986262-3952.09-0.838617.5757583007.947160.893858-3799.39-0.8062110.606063361.557441.652442-3726.85-0.7908267744.149044-3617.75-0.7676778297.985445-3451.69-0.7324389125.962412-3139.66-0.6662299817.617849-2995.82-0.63571010672.1152-2867.52-0.608471112175.03348-2335.53-0.495591213103.82655-1939.63-0.41
11、1581314087.59264-1997.09-0.423771415647.70607-1555.81-0.330131518057.14802-853.848-0.181181622584.88741-684.987-0.145351728999.68001242.520.0514611835279.310451468.890.31169J940567.30083352.1990.7113182044187.840223952.760.8387542146541.874385046.3261.0708032248760.794686876.1051.4590722352443.69806
12、9072.3021.9250932457409.144529524.7552.0211012562939.103168877.3971.8837352670745.080646940.4191.4727192782418.541425134.0591.0894192895138.228873913.0710.83033229112098.1342533.76580.11326230133156.9119-1646.81-0.3494413.636363714.816.666674126.419.696974846.322.727275986.325.757586821.828.78788780
13、4.631.818189839.534.8484811161.237.8787912090.540.9090914091.943.9393917203.346.969721899.95029242.253.030336748.256.0606143919.559.0909148140.662.1212151588.265.1515255636.968.181826151671.2121266933.974.2424271816.577.2727377685.580.3030387552.683.3333399051.386.36361112631.989.39394131510.1311568
14、59.7573-4513.16-O.9576792.42424152346.632172035.4544-5215.35-1.1066795.45455166820.133195218.3264-8313.03-1.7639898.48485186905.3表2-2XVariable1ResidualPlot12000100008000-6000 -4000200050000-2000 -4000 -. 100000150000. Illl200000250000300000350000.400000450000-6000-8000X Variable 1-100OO残差图XVariable1
15、LineFitPlot线性拟合图NormalProbabilityPlot.QbbI020,I406080I100I120SamplePercentile200000180000160000140000120000A100000800006000040000200000正态概率图记最终消费额为X,GDP为Y,则根据表2-1的数据估计回归系数为：_IIEXY-EXEY_33X3.27736X1。113010790.9X1610877.1Bl=nX2-(X)2=336,48898IO11-3010790.92=0.48830670750=-1=48814.45758-0.4883067075X91
16、236.08788=4741.307494所以你和线性回归方程为：Y=4741.307494+0.4883067075X估计标准误差和判定系数：因为e2=Y2-0Y-XY=710692258.5,n=33所以回归标准误710692258.533-2=22925556.72 = 4788.063985判定系数r2=l-=l-。严=1-8.8165XIO-6=0.999991184TSS(yi-y)2对回归系数进行显著性检验由于=0.007827176SBl =则jtg=S15=62.38606291lSBl查t值分布表，显著水平5%,自由度为31的双侧t检验临界值为2,0423tg,所以，没有充
17、分理由接受原假设，可以认为瓦显著不为0,即回归系数Bl对y的影响显著。当2011年我国的国内生产总值为471563.7亿元时，利用拟合的回归方程预测该年可能达到的消费额，给出置信度为95%的预测区间。对于回归方程Y=4741.307494+0.4883067075X,当X。=471563.7时,可得：Y=4741.307494+0.4883067075X471563.7=235009.0252预测标准误差=Sind=SeXjli=4788.063985XJl+表+蔻馨=5699.311538查t值分布表,显著水平5%,自由度为31的双侧t检验临界值为2.0423o因此，当国内生产总值为471563.7亿元时，置信度为95%的预测区间为：235009.0252-2.0423X5699.311538y0235009.0252+2.0423X5699.311538整理得：223369.3212%246648.729即：在概率保证程度为95%的条件下，2011年我国消费额的预测区间为223369.3212（亿元）到246648.729（亿元）之间。
