NMath Stats User's Guide

TOC | Previous | Next | Index

4.3 Box-Cox Power Transformations (.NET, C#, CSharp, VB, Visual Basic, F#)

Box-Cox power transformations compute a rank-preserving transformation of data to stabilize variance and make the data more normal. The power transformation is defined as a continuously varying function, with respect to the power parameter ,

 

In NMath Stats, class BoxCox compute the Box-Cox power tranformations for a set of data points and parameter value . In addition, methods are provided for computing the corresponding log-likelihood function and the value of which maximizes it.

For example:

Code Example – C# Box-Cox transformations

var data = new DoubleVector( "[.15 .09 .18 .10 .05 .12 .08 .05 .08 
.10 .07 .02 .01 .10 .10 .10 .02 .10 .01 .40 .10 .05 .03 .05 .15 .10 
.15 .09 .08 .18 .10 .20 .11 .30 .02 .20 .20 .30 .30 .40 .30 .05]" 
);

var interval = new Interval( -5, 5, Interval.Type.Closed );

var bc = new BoxCox( data, interval );

Console.WriteLine( bc.Lambda );
Console.WriteLine( bc.TransformedData );

BoxCox searches from -5 to 5 until the best value of is found (the value which maximizes the log-likelihood function).


Top

Top