MIXED

MIXED 可在 自訂表格及進階統計量中使用。

MIXED 程序適合各種混合線性模型。 「混合線性模型」會延伸 GLM 程序中使用的廣義線性模型，以使資料能夠展示相關性和非常數變異性。

MIXED dependent varname [BY factor list] [WITH covariate list] 

[/CRITERIA = [CIN({95** })] [HCONVERGE({.00000001**  } {ABSOLUTE  })
                  {value}              {value} {RELATIVE**}

             [DFMETHOD({SATTERTHWAITE**} {RESIDUAL} {KENWARDROGER})]

             [LCONVERGE({0**  } {ABSOLUTE**})] [MXITER({100**})]
                        {value} {RELATIVE  }           {n    }

             [MXSTEP({10**})] [PCONVERGE({0** },{ABSOLUTE**})] [SCORING({1**})]
                     {n  }              {value } {RELATIVE  }            {n  }

             [SINGULAR({1E-12**})] ]
                       {value  } 

[/EMMEANS = TABLES ({OVERALL          })]
                    {factor           }
                    {factor*factor ...}
            [WITH (covariate=value [covariate = value ...])
            [COMPARE [({factor})] [REFCAT({value})] [ADJ({LSD**     })] ]
                                          {FIRST}        {BONFERRONI}
                                          {LAST }        {SIDAK     }

[/FIXED = [effect [effect ...]] [| [NOINT] [SSTYPE({1  })] ] ]
                                                   {3**}

[/METHOD = {ML    }]
           {REML**}

[/MISSING = {EXCLUDE**}]
            {INCLUDE  }

[/PRINT = [CORB] [COVB] [CPS] [DESCRIPTIVES] [G] [HISTORY(1**)] [LMATRIX] [R]
                                                         (n  )
          [SOLUTION] [TESTCOV]]

[/RANDOM = effect [effect ...]
           [| [SUBJECT(varname[*varname[*...]])] [COVTYPE({VC**      })] [SOLUTION]]
                                                          {covstruct+}

[/REGWGT = varname]

[/REPEATED = varname[*varname[*...]] | SUBJECT(varname[*varname[*...]]) KRONECKER(varname[*varname[*...]])
             [COVTYPE({DIAG**    })]]
                      {covstruct†}
             [SPCOORDS(varname...)]

[/SAVE = [tempvar [(name)] [tempvar [(name)]] ...]

[/OUTFILE [EBLUPS('savfile' | dataset)]
[FILE_SEPARATE({FALSE**} {TRUE})]]

[/TEST[(valuelist)] = 
    ['label']  effect valuelist ... [| effect valuelist ...] [divisor=value]]
    [; effect valuelist ... [| effect valuelist ...] [divisor=value]]

[/TEST[(valuelist)] = ['label'] ALL list [| list] [divisor=value]]
                      [; ALL list [| list] [divisor=value]]

** 如果省略次指令，則為預設值。

† covstruct 可採用下列值：AD1、AR1、ARH1、ARMA11、CS、CSH、CSR、DIAG、FA1、FAH1、HF、ID、TP、TPH、UN、UNR、VC。

對於 REPEATED 次指令，covstruct 可採用下列值：SP_POWER、SP_EXPONENTIAL、SP_GAUSSIAN、SP_LINEAR、SP_LINEARLOG、SP_SPHERICAL、UN_AR1、UN_CS、UN_UN。

僅當 COVTYPE 定義為 UN_AR1、UN_CS 或 UN_UN 時才應使用 KRONECKER 關鍵字。

如果 KRONECKER 和 REPEATED 都有效，則它們不一定有一般欄位，但其值不能完全相同（即使是值的順序不同）。

此指令會讀取作用中資料集，並導致執行任何擱置指令。 如需相關資訊，請參閱主題 指令順序 。

可以從「 線性混合模型 」對話框產生 MIXED 指令的語法。

發行歷程

範例

MIXED Y.