ISmLogisticRegression

Assembly: Stat;

Namespace: Prognoz.Platform.Interop.Stat;

Description

The ISmLogisticRegression interface is used for data mining using the Logistic Regression method.

Inheritance Hierarchy

          IStatMethod

          IStatMethodValid

          IDataMining

          ISmLogisticRegression

Comments

This method is used to predict probability of a certain event based on a number of features. For example, if you need to predict whether a certain patient has a disease based on patients' age and gender. This analysis can be used only for binary data.

An explained series must be an array of binary values that indicate the presence or the absence of a characteristic. If observation is set to Double.Nan, it means that observation is skipped and it is required to find whether the characteristic is present. Explained series must represent categorical data.

Properties

  Property name Brief description
The ClassificationSummary property returns summary results of classification.
Outdated. Use IDataMining.FilledDependent.
ModelCoefficients The ModelCoefficients property returns calculated model coefficients.
The MaxIteration property determines the maximum number of iterations, in which the solution must be found.
The NumOfIter property returns the number of iterations within which the decision was found.
Probabilities The Probabilities property returns a series of forecast probabilities of logistic regression.
ProbFitted The ProbFitted property returns probabilities for training objects.
The RelevanceMeasure property returns binary classification quality criteria.
The ROCcurve property returns ROC curve parameters.
SummaryStatistics The SummaryStatistics property returns calculated summary statistics for a model.
The Threshold property determines threshold value of probability for classification.
The Tolerance property determines accuracy of solution.

Properties inherited from IDataMining

  Property name Brief description
The Dependent property returns explained series.
Outdated. Use IDataMining.Dependent.
The Explanatories property returns a collection of classification attributes.
The FilledDependent property returns a series with calculation results.

Properties inherited from IStatMethodValid

  Property name Brief description
CrossValidation The CrossValidation property returns cross-validation settings.
The PerformanceScores property returns cross-validation results.

Properties inherited from IStatMethod

  Property name Brief description

DisplayName

The DisplayName property returns the displayed method name.

ErrorByStatus

The ErrorByStatus property returns an error message by the error number.

Errors

The Errors property returns a message with all the errors and warnings.

Name

The Name property returns the internal method name.

PerformanceTime

The PerformanceTime property returns method execution time.

Status

The Status property returns the method execution status.

SupportsR

The SupportsR property returns whether statistical method can be calculated via R package.

UseR

The UseR property determines whether statistical method is calculated via the R package.

WarningByStatus

The WarningByStatus property returns a warning text by its number.

Warnings

The Warnings property returns the warnings that occurred at method calculation.

WarningsCount

The WarningsCount property returns the number of warnings that occurred at the method calculation.

WarningsNumbers

The WarningsNumbers property returns numbers of warnings that occurred at the method calculation.

Class object methods inherited from IStatMethodValid

  Method name Brief description
The ExecuteValidation method executes cross-validation.

Methods inherited from IStatMethod

  Method Name Brief description

Clone

The Clone method clones a statistical method object.

Execute

The Execute method executes a statistical method.

LoadFromXML

The LoadFromXML method loads statistical method settings from XML code.

SaveToXML

The SaveToXML method unloads statistical method settings to XML code.

See also:

Stat Assembly Interfaces | Logistic regression | Pattern substitution