Threshold: Double;
The Threshold property determines a threshold value.
Range of available values: (0;1].
Executing the example requires that the repository contains:
A table with the DM_TABLE identifier containing data for analysis
A regular report with the DM_REPORT_RES identifier, to which analysis results will be loaded.
Add links to the Metabase, Ms, Report, Stat, Tab system assemblies.
Sub UserProc;
Var
mb: IMetabase;
ReportDS: IDmReportDataSource;
TableDS: IDmTableDataSource;
Method: IDmMethod;
Report: IPrxReport;
Shs: IPrxSheets;
Sheet: ITabSheet;
DM: IDmLogisticRegression;
i, j: Integer;
Attrs: Array Of Integer;
CrossValidation: IDmMethodCrossValidation;
Reports: IDmReports;
DmReport: IDmReport;
CrossValPerf: ICrossValidationPerformanceScores;
CategoriesList: Array Of Integer;
PerformanceMatrix: Array Of Double;
Begin
mb := MetabaseClass.Active;
// Create calculation method
Method := (New DataMiningMethod.Create) As IDmMethod;
// Specify method type
Method.Kind := DmMethodKind.LogisticRegression;
// Create table data source
TableDS := (New TableDataSource.Create) As IDmTableDataSource;
// Determine source table
TableDS.Table := mb.ItemByID("DM_TABLE").Bind;
// Set input data source
Method.InputDataSource := TableDS;
// Create a data source that is a regular report
ReportDS := (New ReportDataSource.Create) As IDmReportDataSource;
// Set data consumer
Method.OutputDataSource := ReportDS;
// Set up calculation method parameters
DM := Method.Details As IDmLogisticRegression;
// Set factors impacting analyzed attribute
Attrs := New Integer[TableDS.FieldCount - 1];
For i := 0 To Attrs.Length - 1 Do
Attrs[i] := i + 1;
End For;
DM.Attributes := Attrs;
// Specify analyzed object
DM.Target := ReportDS.FieldCount;
// Set threshold
DM.Threshold := 0.57;
// Set up cross-validation parameters
CrossValidation := DM.CrossValidation;
// Specify that cross-validation is used
CrossValidation.Active := True;
// Set cross-validation method: Repeated random sub-sampling cross-validation
CrossValidation.SamplingType := CrossValidationSamplingType.RandomSampling;
// Set number of repetitions
CrossValidation.NumberOfRandomTests := 15;
// Set training set size
CrossValidation.TrainingSetSize := 75;
// Perform analysis and output results
Reports := Method.Execute;
DmReport := reports.FindByType(DmReportType.LogisticRegression);
ReportDS := DmReport.Generate;
ReportDS.TabSheet.View.Selection.SelectAll;
ReportDS.TabSheet.View.Selection.Copy;
// Get regular report, to which results will be loaded
Report := mb.ItemByID("DM_REPORT_RES").Edit As IPrxReport;
Shs := Report.Sheets;
Shs.Clear;
Sheet := (Shs.Add("", PrxSheetType.Table) As IPrxTable).TabSheet;
Sheet.Table.Paste;
Sheet.Columns(0, 1).AdjustWidth;
Sheet.Rows(0, 1).AdjustHeight;
Report.Sheets.Item(0).Name := ReportDS.Caption;
// Save loaded data
(Report As IMetabaseObject).Save;
// Display cross-validation results
DmReport := reports.FindByType(DmReportType.CrossValidation);
DmReport.Generate;
CrossValPerf := CrossValidation.Results;
Debug.WriteLine("Cross-validation results:");
Debug.Indent;
Debug.WriteLine("Analyzed attribute: " + CrossValPerf.ClassificatorName);
Debug.Write("Number of factors influencing the analyzed attribute: ");
Debug.WriteLine(CrossValPerf.FactorsNumber);
Debug.WriteLine("Number of observations: " + CrossValPerf.ObservationsNumber.ToString);
Debug.WriteLine("Number of repetitions: " + CrossValidation.NumberOfRandomTests.ToString);
Debug.WriteLine("Accuracy of classification: " + CrossValPerf.ClassificationAccuracy.ToString);
Debug.WriteLine("Categories:");
Debug.Indent;
CategoriesList := CrossValPerf.CategoriesList;
For i := 0 To CategoriesList.Length - 1 Do
Debug.WriteLine(CategoriesList[i]);
End For;
Debug.Unindent;
Debug.WriteLine("Correct classification:");
Debug.Indent;
PerformanceMatrix := CrossValPerf.PerformanceMatrix;
For i := 0 To PerformanceMatrix.GetUpperBound(1) Do
For j := 0 To PerformanceMatrix.GetUpperBound(2) Do
Debug.Write(PerformanceMatrix[i, j].ToString + #9);
End For;
Debug.WriteLine("");
End For;
Debug.Unindent;
Debug.Unindent;
End Sub UserProc;
After executing the example, the missing data substitution using logistic regression will be executed for the DM_TABLE table data. Analysis results will be loaded to the DM_REPORT_RES report. Cross-validation results will be displayed in the console window.
See also: