影响房价的多元线性回归.pptx

影响房价的多元线性回归分析数据____数据来源于国家统计局官网，年度数据年份全国平均房价（元/平米）国民收入（亿元)商品房销售面积（万平米）建筑企业个数（个）居民消费额（亿元)轿车产量（万辆）家用电冰箱产量（万台）19982063301012185.345,634188850.711,060.0019992053335014556.547,234219557.11,210.0020002058368818637.147,518253160.71,279.0020012170394122411.945,893279970.361,351.2620022250425826808.347,8203054109.21,598.8720032359473633717.648,6883358207.082,242.5620042714565238231.659,0183905227.633,007.592005316770205548658,7504879277.012,987.062006336778596185760,1665552386.943,530.892007386493137735562,0746386479.784,397.1320083800117386597071,0958038503.814,799.9520094681131769475570,8179005748.485,930.45201050321438410476571,8639663957.597,295.72201153771655710994672,280109691,012.678,699.20201257912022311130075,280129851,077.008,427.00201362372488213055179,528159491,210.429,261.02Excel进行数据分析后的结果：人工分析1回归统计MultipleR0.998916RSquare0.997833AdjustedRSquare0.996389标准误差87.47509观测值16（1）MultipleR相关系数趋近1，说明变量间的线性关系密切程度大，呈正相关；（2）RSquare可决系数和AdjustedRSquare调整后的可决系数都是0.99几，所以我们对回归模型以非常肯定的态度；人工分析2　系数R相关性数t检验值显著水平F值下限95.0%上限95.0%截距907998916690.8051X10.3304551852.423764010.038373020.022033420.638876945X20.0199741864.0914236560.002711260.008930410.031017958X30.0164118361.81655236400036849536X4-0.498858792-2.09725570201039223897X50.9892902921.30064589700709920112X60.0056747340.06222591900211973656(1)这里取小数点后4为，四舍五入得多元线性回归方程：Y=0.3305x1+0.1997x2+0.0164x3-0.4989x4+0.9892x5+0.0057x6+907.8298(2)蓝色部分的是置信区间，上下限可以理解为变量x的取值范围.人工分析3　R相关性数t检验值显著水平F值下限95.0%上限95.0%X10.9989162.423764010.03837302690.80510.022033420.638876945X24.0914236560.002711260.008930410.031017958X31.81655236400036849536X4-2.09725570201039223897X51.30064589700709920112X60.06222591900211973656显著水平的0.05和0.01是通常使用的值，可以理解为国际惯例，我在这里选择0.05这里标注红色的数据都小于0.05，说明x1，x2两个变量的变化对y的影响明显，即说明国民收入和商品房销售面积对房价的影响明显。人工分析4F值690.8051给定α查F分布表,得临界值Fα(p,n-p-1)。其中α为显著性水平（0.05），n（16）为样本个数，p（6