 | BoxCox Class |
Class for computing the Box-Cox power tranformations defined for a set of data
points, {yi}, and parameter value lambda by
yi(lambda) = (yi^lambda - 1)/lambda.
In addition methods for computing the corresponding log-likelihood function and
the value of lambda which maximizes it are provided.
Inheritance HierarchyNamespace: CenterSpace.NMath.Core
Assembly: NMath (in NMath.dll) Version: 7.3
SyntaxThe BoxCox type exposes the following members.
Constructors| Name | Description | |
|---|---|---|
 | BoxCox(DoubleVector) |
Constructs a BoxCox object for the given data, finding a value
of lambda which maximizes the log-likelihood function over the lambda
interval LAMBDA_EPSILON.
|
| BoxCox(DoubleVector, Interval) |
Constructs a BoxCox object for the given data. A value for
lambda which maximizes the log-likelihood function for the given
interval for lambda is computed.
|
| BoxCox(DoubleVector, Interval, Double) |
Constructs a BoxCox object for the given data. A value for
lambda which maximizes the log-likelihood function for the given
interval for lambda is computed.
|
| BoxCox(DoubleVector, Interval, Double, Double) |
Constructs a BoxCox object for the given data. A value for
lambda which maximizes the log-likelihood function for the given
interval for lambda is computed. This maximum value is found by
manually searching the lambda values lambdaInterval.Min + k*stepSize,
k = 0, 1, 2,... and selecting the one which yields the maximum value
for the log-likelihood function.
|
| BoxCox(DoubleVector, Double, Double, Double) |
Constructs a BoxCox object for the given data. A value for
lambda which maximizes the log-likelihood function for the given
interval for lambda is computed. This maximum value is found by
manually searching the lambda values minLambda + k*stepSize,
k = 0, 1, 2,... and selecting the one which yields the maximum value
for the log-likelihood function.
|
Properties| Name | Description | |
|---|---|---|
 | Lambda |
Gets the value of lambda which maximizes the log-likelihood function
for the data and lambda interval.
|
| LambdaEpsilon |
Lambda values less than LambdaEpsilon were considered equal
to zero.
|
| LambdaInterval |
The interval that was searched for the lambda value that
maximized the log-likelihood function.
|
| TransformedData |
The data transformed by the Box Cox transformation
yi(lambda) = (yi^lambda - 1)/lambda
where lambda is the value of lambda the
minimized the log-likelihood function of the sprecified
search interval LambdaInterval.
|
Methods| Name | Description | |
|---|---|---|
 | LogLikelihood(DoubleVector, Double) |
Calculates the Box Cox log-likelihood function for the given data and
and lambda value.
|
| LogLikelihood(DoubleVector, Double, Double) |
Calculates the Box Cox log-likelihood function for the given data and
and lambda value.
|
| Transform(DoubleVector, Double) |
Performs the Box Cox transformation
transformed data[i] = (data[i]^lambda - 1)/lambda.
|
| Transform(DoubleVector, Double, Double) |
Performs the Box Cox transformation
transformed data[i] = (data[i]^lambda - 1)/lambda.
|
Fields| Name | Description | |
|---|---|---|
 | LAMBDA_EPSILON |
Default value for lambdaEpsilon. Values for lambda less than lambdaEpsilon
are considered to be zero.
|
| LAMBDA_INTERVAL |
The default interval to search for a value of lambda maximizing the
log-likelihood functions. The default interval is [-5, 5].
|
See Also