《SPSS处理多元判别分析》由会员分享，可在线阅读，更多相关《SPSS处理多元判别分析（12页珍藏版）》请在金锄头文库上搜索。
1、实验四：多元判别分析实验目的1熟练掌握使用SPSS对数据进行多元判别分析的方法掌握对数据的多元判别结果的分析方法实验要求能够按照实验题目要求完成实验题目掌握实验中要求的掌握方法，熟练操作SPSS对实验结果进行分析实验内容实验过程：1依次点击“分析一一分类一一判别”并设置相关量，如下图一所示:【图一】分析的结果如表一所示：【表一】判别分析案例处理摘要未加权案例N百分比有效100100.0排除的缺失或越界组代码0.0至少一个缺失判别变量0.0缺失或越界组代码还有至少一个缺失判别变量0.0合计0.0合计100100.0组统计量group均值标准差有效的N（列表状态）未加权的已加权的正常人pa28.2
2、1364.700562525.000汇聚的组内矩阵paalpha_aghpalpha_at相关性pa1.000-.112.119-.290alpha_ag-.1121.000.456.528hp.119.4561.000.484alpha_at-.290.528.4841.000肝硬化合计alpha_ag67.578016.752412525.000hp257.1212126.276842525.000alpha_at282.168030.833372525.000pa15.855510.210724040.000alpha_ag120.794362.047904040.000hp321.83
3、57249.334074040.000alpha_at492.4633151.322534040.000pa16.31457.801522020.000alpha_ag55.298026.128322020.000hp91.4700126.450502020.000alpha_at313.308055.596232020.000pa21.97938.472641515.000alpha_ag69.618750.464771515.000hp297.1527210.051231515.000alpha_at314.728772.527361515.000pa19.95549.7761210010
4、0.000alpha_ag86.714653.67732100100.000hp255.8815212.46384100100.000alpha_at377.3982140.18786100100.000肝癌，AFP检测阴性肝癌，AFP检测阳性分析1协方差矩阵的均等性的箱式检验对数行列式group秩对数行列式正常人425.055肝癌，AFP检测阳性432.930肝癌，AFP检测阴性426.634肝硬化429.759汇聚的组内430.930打印的行列式的秩和自然对数是组协方差矩阵的秩和自然对数检验结果箱的M160.976F近似。4.888df130df211927.756Sig.000对相等总体
5、协方差矩阵的零假设进行检验。特征值函数特征值方差的%累积%正则相关性1.980(a)76.576.5.7032.260(a)20.396.7.4543.042(a)3.3100.0.200a分析中使用了前3个规范判别式函数Wilks的Lambda函数检验Wilks的Lambda卡方dfSig.1到3.38590.71912.0002到3.76225.8416.0003.9603.8812.144标准化的规范判别式函数系数1函数23pa-.203.691.765alpha_ag.221.303.322hp-.312.471-1.106alpha_at.898.103.473结构矩阵函数123al
6、pha_at.922(*).290-.114alpha_ag.576(*).494-.018hp.199.740(*)-.640pa-.525.683(*).460*每个变量和任意判别式函数间最大的绝对相关性规范判别式函数系数函数123pa-.024.082.091alpha_ag.005.007.007hp-.002.002-.006alpha_at.009.001.005（常量）-2.772-3.193-2.696非标准化系数组质心处的函数group函数123正常人-1.112.464.181肝癌，AFP检测阳性1.146.156.014肝癌，AFP检测阴性-.353-.961.082肝硬
7、化-.735.092-.450在组均值处评估的非标准化规范判别式函数分类统计量分类处理摘要已处理的100已排除的缺失或越界组代码0至少一个缺失判别0变量用于输出中100group先验用于新谢案例未加权的已加权的正窜人肝癌，亞届测阳性肝廈j测阴性肝硬化合计2S02502502501.0002S4020151C02S0004000020,000150001C0000区域图（假定前两个函数以外的所有函数为0）典则判别函数2-6.0-4.0-2.0.02.04.06.06.01212121212124.01212121212122.012121121442*1144212.01444444321*44
8、42*1113333*332111133332111333332-2.01113333221113333321113333211133332333322332-4.0323232233232-6.03211443333332232-6.0-4.0-2.0.02.04.06.0典则判别函数1区域图中使用的符号符号组标签1正常人2肝癌，AFP佥测阳性3肝癌，AFP佥测阴性4肝硬化*表示一个组质心按照案例顺序的统计量案例数目实际组最高组组预测组P(Dd|G=g)P(G=g|D=d)到质心的平方Mahalanobis距离Pdf初始113(*).9573.558.3174211.9383.374.412
9、4314(*).7383.5521.2631411.9813.399.1774511.9073.412.5524611.9283.387.4584711.8713.398.7113814(*).9453.445.3761911.9903.488.11341011.9733.428.22641111.9863.450.14241214(*).6663.5011.57311311.5603.4822.06031411.8613.499.75441511.9243.507.47741611.9903.495.11541711.9913.516.10541814(*).4053.4812.914119
10、11.8933.512.61342011.6843.5681.49342111.9623.589.29142211.9503.603.35442311.8513.638.79542411.8503.620.79942511.7253.7171.31542623(*).2873.8003.77142723(*).5133.7712.30022823(*).3933.6852.99222922.4663.7732.55433022.00131.00016.34733122.0933.9456.40933222.2853.9293.78943322.4393.8022.70533422.6253.6271.75433523(*).7443.6071.23823622.6143.9511.80533723(*).6383.4231.69723822.3573.9673.23133922.0103.98711.45344022.0883.9856.53834122.5573.9712.07734223(*).8363.698.85744322.9713.614.23744422.9553.730.32534522.8873
