Instrumental: ISlSeries;
Instrumental: Prognoz.Platform.Interop.Stat.ISlSeries;
The Instrumental property determines instrumental variables.
To determine whether the constant is used as an instrumental variable, use the ISm2SLS.UseConstantAsInstrument property.
To execute the example, add a link to the Stat system assembly.
Sub UserProc;
Var
TwoSLS: Sm2SLS;
can, fra, ger, jpn, uk: Array[15] Of Double;
C: IIntercept;
MC: ICoefficients;
X: ISlSerie;
res,i: Integer;
Begin
// Set values for variables
can[00] := Double.Nan; fra[00] := 4110; ger[00] := 3415; jpn[00] := 1475; uk[00] := 5320;
can[01] := 6385; fra[01] := 4280; ger[01] := 3673; jpn[01] := 1649; uk[01] := 5484;
can[02] := 6752; fra[02] := 4459; ger[02] := 4013; jpn[02] := Double.Nan; uk[02] := 5517;
can[03] := 6837; fra[03] := 4545; ger[03] := 4278; jpn[03] := 1884; uk[03] := 5791;
can[04] := 6495; fra[04] := 4664; ger[04] := 4577; jpn[04] := 1972; uk[04] := 5971;
can[05] := 6907; fra[05] := Double.Nan; ger[05] := 5135; jpn[05] := 2108; uk[05] := Double.Nan;
can[06] := 7349; fra[06] := 5195; ger[06] := 5388; jpn[06] := 2249; uk[06] := 6238;
can[07] := 7213; fra[07] := 5389; ger[07] := 5610; jpn[07] := 2394; uk[07] := 6322;
can[08] := 7061; fra[08] := 5463; ger[08] := Double.Nan; jpn[08] := 2505; uk[08] := 6340;
can[09] := 7180; fra[09] := 5610; ger[09] := 6181; jpn[09] := 2714; uk[09] := 6569;
can[10] := 7132; fra[10] := 5948; ger[10] := 6633; jpn[10] := 3052; uk[10] := 6813;
can[11] := Double.Nan; fra[11] := 6218; ger[11] := 6910; jpn[11] := 3453; uk[11] := 6974;
can[12] := 7473; fra[12] := 6521; ger[12] := 7146; jpn[12] := 3666; uk[12] := 6994;
can[13] := 7722; fra[13] := 6788; ger[13] := 7248; jpn[13] := 4008; uk[13] := 7220;
can[14] := 8088; fra[14] := 7222; ger[14] := 7689; jpn[14] := 4486; uk[14] := Double.Nan;
// Create model
TwoSLS := New Sm2SLS.Create;
// Set sample period parameters
TwoSLS.ModelPeriod.FirstPoint := 1;
TwoSLS.ModelPeriod.LastPoint := 10;
// Set forecast period parameters
TwoSLS.Forecast.LastPoint := 15;
// Use auto estimation of constant value
C:=TwoSLS.ModelCoefficients.Intercept;
C.Mode := InterceptMode.AutoEstimate;
// Set explained variable
TwoSLS.Explained.Value := can;
// Set explanatory variables
X:=TwoSLS.Explanatories.Add;
X.Value := fra;
X.Name:= "fra";
X:=TwoSLS.Explanatories.Add;
X.Value := ger;
X.Name:= "ger";
// Use constants in instrumental variables
TwoSLS.UseConstantAsInstrument := True;
// Set instrumental variables
TwoSLS.Instrumental.Add.Value := jpn;
TwoSLS.Instrumental.Add.Value := uk;
// Set missing data treatment method
TwoSLS.MissingData.Method := MissingDataMethod.AnyValue;
// Run calculation and show results
res := TwoSLS.Execute;
If res <> 0 Then
Debug.WriteLine(TwoSLS.Errors);
Else
Debug.WriteLine("=== Model coefficients ===");
Debug.WriteLine("Constant: " + C.Estimate.ToString);
MC := TwoSLS.ModelCoefficients.Coefficients;
For i := 0 To MC.Estimate.Length - 1 Do
Debug.WriteLine(TwoSLS.Explanatories.Item(i).Name + ": " + MC.Estimate[i].ToString);
End For;
Debug.WriteLine(" === Descriptive statistics === ");
Debug.WriteLine("Determination coefficient: " + TwoSLS.SummaryStatistics.R2.ToString);
Debug.WriteLine("Sum of residuals squares: " + TwoSLS.SummaryStatistics.SSR.ToString);
Debug.WriteLine("Standard regression error: " + TwoSLS.SummaryStatistics.SE.ToString);
Debug.WriteLine("");
Debug.WriteLine(" === Model residuals === ");
For i := 0 To TwoSLS.Residuals.Length - 1 Do
Debug.WriteLine(i.ToString + " " + TwoSLS.Residuals[i].ToString);
End For;
Debug.WriteLine(" === Modeling series === ");
For i := 0 To TwoSLS.Fitted.Length - 1 Do
Debug.WriteLine(i.ToString + " " + TwoSLS.Fitted[i].ToString);
End For;
Debug.WriteLine(" === Forecast series === ");
For i := 10 To TwoSLS.Forecast.Value.Length - 1 Do
Debug.WriteLine(i.ToString + " " + TwoSLS.Forecast.Value[i].ToString);
End For;
End If;
End Sub UserProc;
Imports Prognoz.Platform.Interop.Stat;
…
Public Shared Sub Main(Params: StartParams);
Var
TwoSLS: Sm2SLS;
can, fra, ger, jpn, uk: Array[15] Of Double;
C: IIntercept;
MC: ICoefficients;
X: ISlSerie;
res,i: Integer;
Estimate, Residuals, Forecast: System.Array;
Begin
// Set values for variables
can[00] := Double.Nan; fra[00] := 4110; ger[00] := 3415; jpn[00] := 1475; uk[00] := 5320;
can[01] := 6385; fra[01] := 4280; ger[01] := 3673; jpn[01] := 1649; uk[01] := 5484;
can[02] := 6752; fra[02] := 4459; ger[02] := 4013; jpn[02] := Double.Nan; uk[02] := 5517;
can[03] := 6837; fra[03] := 4545; ger[03] := 4278; jpn[03] := 1884; uk[03] := 5791;
can[04] := 6495; fra[04] := 4664; ger[04] := 4577; jpn[04] := 1972; uk[04] := 5971;
can[05] := 6907; fra[05] := Double.Nan; ger[05] := 5135; jpn[05] := 2108; uk[05] := Double.Nan;
can[06] := 7349; fra[06] := 5195; ger[06] := 5388; jpn[06] := 2249; uk[06] := 6238;
can[07] := 7213; fra[07] := 5389; ger[07] := 5610; jpn[07] := 2394; uk[07] := 6322;
can[08] := 7061; fra[08] := 5463; ger[08] := Double.Nan; jpn[08] := 2505; uk[08] := 6340;
can[09] := 7180; fra[09] := 5610; ger[09] := 6181; jpn[09] := 2714; uk[09] := 6569;
can[10] := 7132; fra[10] := 5948; ger[10] := 6633; jpn[10] := 3052; uk[10] := 6813;
can[11] := Double.Nan; fra[11] := 6218; ger[11] := 6910; jpn[11] := 3453; uk[11] := 6974;
can[12] := 7473; fra[12] := 6521; ger[12] := 7146; jpn[12] := 3666; uk[12] := 6994;
can[13] := 7722; fra[13] := 6788; ger[13] := 7248; jpn[13] := 4008; uk[13] := 7220;
can[14] := 8088; fra[14] := 7222; ger[14] := 7689; jpn[14] := 4486; uk[14] := Double.Nan;
// Create model
TwoSLS := New Sm2SLS.Create();
// Set sample period parameters
TwoSLS.ModelPeriod.FirstPoint := 1;
TwoSLS.ModelPeriod.LastPoint := 10;
// Set last forecast point
TwoSLS.Forecast.LastPoint := 15;
// Use auto estimation of constant value
C:=TwoSLS.ModelCoefficients.Intercept;
C.Mode := InterceptMode.imAutoEstimate;
// Set explained variable
TwoSLS.Explained.Value := can;
// Set explanatory variables
X:=TwoSLS.Explanatories.Add();
X.Value := fra;
X.Name:= "fra";
X:=TwoSLS.Explanatories.Add();
X.Value := ger;
X.Name:= "ger";
// Use constants in instrumental variables
TwoSLS.UseConstantAsInstrument := True;
// Set instrumental variables
TwoSLS.Instrumental.Add().Value := jpn;
TwoSLS.Instrumental.Add().Value := uk;
// Set missing data treatment method
TwoSLS.MissingData.Method := MissingDataMethod.mdmAnyValue;
// Run calculation and show results
res := TwoSLS.Execute();
If res <> 0 Then
System.Diagnostics.Debug.WriteLine(TwoSLS.Errors);
Else
System.Diagnostics.Debug.WriteLine(" === Model coefficients ===");
System.Diagnostics.Debug.WriteLine("Constant: " + C.Estimate.ToString());
MC:=TwoSLS.ModelCoefficients.Coefficients;
Estimate := MC.Estimate;
For i:=0 To MC.Estimate.Length-1 Do
System.Diagnostics.Debug.WriteLine(TwoSLS.Explanatories.Item[i].Name+": "+Estimate[i].ToString());
End For;
System.Diagnostics.Debug.WriteLine(" === Descriptive statistics === ");
System.Diagnostics.Debug.WriteLine("Determination coefficient: " + TwoSLS.SummaryStatistics.R2.ToString());
System.Diagnostics.Debug.WriteLine("Sum of residual squares: "+TwoSLS.SummaryStatistics.SSR.ToString());
System.Diagnostics.Debug.WriteLine("Standard regression error: "+TwoSLS.SummaryStatistics.SE.ToString());
System.Diagnostics.Debug.WriteLine("");
System.Diagnostics.Debug.WriteLine(" === Model residuals === ");
Residuals := TwoSLS.Residuals;
For i := 0 To TwoSLS.Residuals.Length - 1 Do
System.Diagnostics.Debug.WriteLine(i.ToString() + " " + Residuals[i].ToString());
End For;
System.Diagnostics.Debug.WriteLine(" === Modeling series === ");
Fitted := TwoSLS.Fitted;
For i := 0 To TwoSLS.Fitted.Length - 1 Do
System.Diagnostics.Debug.WriteLine(i.ToString() + " " + Fitted[i].ToString());
End For;
System.Diagnostics.Debug.WriteLine(" === Forecast series === ");
Forecast := TwoSLS.Forecast.Value;
For i:=10 To TwoSLS.Forecast.Value.Length-1 Do
System.Diagnostics.Debug.WriteLine(i.ToString() + " " + Forecast[i].ToString());
End For;
End If;
End Sub;
After executing the example the specific linear regression model is created and calculated. The two-step least squares method is used to estimate coefficients. Calculation results are displayed to the console window.
See also: