hfl109
请教专家:
下面5列数据,前4列是自变量,最后是因变量,怎么分析4个因子对因变量的贡献?数据不是正态分布,怎么可以转化为正态分布?非常感谢
0 0 0 0 99.65
0 0 0 0 99.88
421.14 0 0 0 93.88
342.62 0 0 0 96.09
2727.49 54.6 551.51 46 51.59
3459.48 71.48 637.16 40.47 49.42
264.82 5.94 6.06 0 92.64
224.54 1.68 5.96 0 91.08
0 0 0 0 99.21
0 0 0 0 99.12
483.07 69.26 277.7 27.55 69.44
484.79 69.06 275.65 27.42 69.35
528.49 160.4 33.57 197.59 48.3
607.17 168.16 35.15 214.77 44.87
2231.51 0 33.41 0 79.72
2912.42 29.94 18.08 0 81.96
3212.94 68.57 16.77 0 81.76
1745.29 4.84 16.06 0 85.78
4445.72 142.69 986.15 84.21 34.08
5035.93 153.97 889.09 78.06 31.94
776.54 42.29 5.88 17.55 73.05
534.87 28.45 3.63 9.71 78.95
1.403 0 0 0 99.21
1.571 0 0 0 98.32
阿娅
不知道可不可以说两句话。
一提到贡献 就想到因子分析了,刚在sas中做的,看下有没有什么有用的信息
Eigenvalues of the Correlation Matrix: Total = 4 Average = 1
Eigenvalue Difference Proportion Cumulative
1 2.64369125 1.54271742 0.6609 0.6609
2 1.10097383 0.89057922 0.2752 0.9362
3 0.21039461 0.16545431 0.0526 0.9888
4 0.04494031 0.0112 1.0000
Statsfu
make some transformation like inverse , log. usually they will improve the performance of regression. Or, you can try generalized linear model.
hfl109
非常感谢!
因子分析是分析各因子之间的吧,分析是不是只能对前4列分析,分析出来的贡献不知道是不是可以说就是对因变量y的贡献?
hfl109
[quote]引用第2楼Statsfu于2007-09-18 17:20发表的“”:
make some transformation like inverse , log. usually they will improve the performance of regression. Or, you can try generalized linear model. [/quote]
谢谢!!!
分析看看
hfl109
转化后,也不成正态分布