PROXSCAL

PROXSCAL 在 取樣及檢定中可用。

PROXSCAL 會執行多維度近似性資料的尺度，以尋找低維度空間中個體的最小平方表示法。 多個來源容許個別差異模型。 Majorization 演算法可保證在各種模型和限制下，選擇性地轉換度量和非度量資料的單調收斂。

PROXSCAL varlist

[/TABLE = {rowid BY columnid [BY sourceid]}]
          {sourceid                      }

[/SHAPE = [{LOWER**}]] 
           {UPPER  }
           {BOTH   }

[/INITIAL = [{SIMPLEX**        }]] 
             {TORGERSON        } 
             {RANDOM[({1})]    }
                      {n} 
             {[('file'|'dataset')] [varlist] }

[/WEIGHTS = varlist]

[/CONDITION = [{MATRIX**      }]] 
               {UNCONDITIONAL }

[/TRANSFORMATION = [{RATIO**                            }]] 
                    {INTERVAL                           }
                    {ORDINAL[({UNTIE   })]              }
                              {KEEPTIES}
                    {SPLINE [DEGREE = {2}] [INKNOT = {1}]}
                                      {n}            {n}

[/PROXIMITIES = [{DISSIMILARITIES**}]]
                 {SIMILARITIES     }

[/MODEL = [{IDENTITY**    }]] 
           {WEIGHTED      }
           {GENERALIZED   }
           {REDUCED[({2})]}
                     {n}

[/RESTRICTIONS = {COORDINATES('file'|'dataset') [{ALL    }]                                  }]
                                     {varlist}
                 {VARIABLES('file'|'dataset') [{ALL    }][({INTERVAL                      })]}
                                   {varlist}   {NOMINAL                       }
                                               {ORDINAL[({UNTIE   })]         }
                                                         {KEEPTIES}
                                               {SPLINE[DEGREE={2}][INKNOT={1}]}
                                                              {n}         {n}

[/ACCELERATION = NONE]

[/CRITERIA = [DIMENSIONS({2**      })]
                         {min[,max]}
             [MAXITER({100**})]
                      {n    }
             [DIFFSTRESS({0.0001**})] 
                         {value   } 
             [MINSTRESS({0.0001**}) ]] 
                        {value   }

[/PRINT = [NONE][INPUT][RANDOM][HISTORY][STRESS**][DECOMPOSITION]
          [COMMON**][DISTANCES][WEIGHTS**][INDIVIDUAL]
          [TRANSFORMATIONS][VARIABLES**][CORRELATIONS**]]

[/PLOT = [NONE][STRESS][COMMON**][WEIGHTS**][CORRELATIONS**]
         [INDIVIDUAL({varlist})]
                     {ALL    }
         [TRANSFORMATIONS({varlist}) [({varlist})[...]] ]
                          {ALL    }    {ALL    }
         [RESIDUALS({varlist}) [({varlist})[...]] ]
                    {ALL    }    {ALL    }
         [VARIABLES({varlist})]]
                    {ALL    }

[/OUTFILE = [COMMON('file'|'dataset')] [WEIGHTS('file'|'dataset')] [DISTANCES('file'|'dataset')]
            [TRANSFORMATIONS('file'|'dataset')] [VARIABLES('file'|'dataset')] ]

[/MATRIX = IN('file'|'dataset')]].

** 如果省略次指令，則為預設值。

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 。

可從 多維度方法 (PROXSCAL) 對話框產生 PROXSCAL 指令的語法。