有没有一个对数据集中所有变量进行PCA的命令，而不需要在STATA中一一选择每个变量

Question

我正在 Stata 中对一个包含 20 多个变量的数据集运行 PCA，我想对除 ID 列之外的所有变量执行 PCA。我正在使用 PCA 命令并一一选择所有变量，但遇到了这个错误 - “观察不足”

Answer 1

您可以使用

ds, not

获取变量列表。这对汽车数据集中的所有变量运行 PCA 除了 make:

. sysuse auto, clear
(1978 automobile data)

. ds make, not
price         mpg           rep78         headroom      trunk         weight        length        turn          displacement  gear_ratio    foreign

. pca `r(varlist)'

Principal components/correlation                 Number of obs    =         69
                                                 Number of comp.  =         11
                                                 Trace            =         11
    Rotation: (unrotated = principal)            Rho              =     1.0000

    --------------------------------------------------------------------------
       Component |   Eigenvalue   Difference         Proportion   Cumulative
    -------------+------------------------------------------------------------
           Comp1 |      6.79324      5.46883             0.6176       0.6176
           Comp2 |      1.32441       .30291             0.1204       0.7380
           Comp3 |       1.0215      .508122             0.0929       0.8308
           Comp4 |      .513382      .100772             0.0467       0.8775
           Comp5 |       .41261      .119505             0.0375       0.9150
           Comp6 |      .293105     .0368205             0.0266       0.9417
           Comp7 |      .256285      .101664             0.0233       0.9650
           Comp8 |      .154621     .0290009             0.0141       0.9790
           Comp9 |       .12562     .0474784             0.0114       0.9904
          Comp10 |     .0781414     .0510665             0.0071       0.9975
          Comp11 |     .0270749            .             0.0025       1.0000
    --------------------------------------------------------------------------

Principal components (eigenvectors) 

    ------------------------------------------------------------------------------------------------------------------------------------------
        Variable |    Comp1     Comp2     Comp3     Comp4     Comp5     Comp6     Comp7     Comp8     Comp9    Comp10    Comp11 | Unexplained 
    -------------+--------------------------------------------------------------------------------------------------------------+-------------
           price |   0.1835    0.5458   -0.4885    0.1144   -0.4743    0.2108    0.0601   -0.2901   -0.2067   -0.0933   -0.0897 |           0 
             mpg |  -0.3212   -0.0930    0.1234   -0.3734   -0.3655    0.5508    0.3699    0.3308    0.0971   -0.2012    0.0194 |           0 
           rep78 |  -0.1900    0.5922    0.1615   -0.6208    0.3485   -0.1553    0.1048   -0.1592   -0.0844    0.1126   -0.0014 |           0 
        headroom |   0.2190    0.1345    0.6954    0.1418   -0.4623   -0.3501    0.2767   -0.0996   -0.0675   -0.0546    0.0012 |           0 
           trunk |   0.2838    0.2425    0.4326    0.0782    0.0974    0.5786   -0.5421    0.0985   -0.0844    0.0923    0.0619 |           0 
          weight |   0.3696    0.0843   -0.1034   -0.0464    0.0912    0.0301    0.2026   -0.0230    0.3783   -0.0953    0.8003 |           0 
          length |   0.3614    0.0892    0.0338    0.0211    0.3077    0.0810    0.1814   -0.0264    0.4189   -0.5314   -0.5168 |           0 
            turn |   0.3444   -0.1055   -0.0302    0.0664    0.3367    0.1419    0.4825    0.2099   -0.6727    0.0429    0.0153 |           0 
    displacement |   0.3601    0.0631   -0.1267   -0.1455   -0.1471   -0.0450    0.1232    0.3666    0.3234    0.6915   -0.2705 |           0 
      gear_ratio |  -0.3254    0.1044    0.1370    0.4778    0.2248    0.3230    0.3951   -0.3736    0.2282    0.3642   -0.0283 |           0 
         foreign |  -0.2777    0.4740   -0.0416    0.4236    0.0912   -0.1915    0.0051    0.6650    0.0357   -0.1501    0.0774 |           0 
    ------------------------------------------------------------------------------------------------------------------------------------------

您的错误确实表明其他问题给您带来了麻烦，但如果没有其他详细信息（变量数量、观察结果和缺失数据的模式），很难诊断出问题所在。