>  Docs Center  >  IDL Reference  >  Machine Learning  >  IDLmlTanHNormalizer

IDLmlTanHNormalizer

IDLmlTanHNormalizer

The IDLmlTanHNormalizer class implements a Hyperbolic Tangent Normalizer which maps the data to the Tanh of the data. This spreads out or increases the resolution around the mean and squishes the outliers to the edges. For example, data that follows a normal distribution will concentrate on the middle. The normalized data will be confined to the range (-1, +1).

Example


Data = findgen(10)
Normalizer = IDLmlTanHNormalizer(data)
Print, Normalizer.Normalize(data)

Syntax


Result = IDLmlTanHNormalizer(Data)

Arguments


Data

A numerical array from which the normalizer will determine the proper normalization parameters. The array must be 2D, of size n x m, where n is the number of attributes and m is the number of examples.

Keywords


None

Properties


None

Methods


IDLmlTanHNormalizer::Normalize


The IDLmlTanHNormalizer::Normalize method computes the normalization of input data.

Syntax


This method can be invoked as a procedure or a function. The procedure syntax is faster and more efficient, but it overwrites the Data argument. The function syntax is less efficient, but it leaves the input argument untouched.

As a procedure:

Obj->[IDLmlTanHNormalizer::]Normalize, Data

As a function:

Result = Obj->[IDLmlTanHNormalizer::]Normalize(Data)

Return Value


If called as a procedure, the normalized array will overwrite the input Data argument. If called as a function, the return value will be an array with the same dimensions as Data, which contains the normalized result.

Arguments


Data

Specify a scalar, 1D array or 2D array of data to normalize.

If the data is 2D, the array must be of size n x m, where n is the number of attributes and m is the number of examples. If the data is 1D, the array must be of size n (number of attributes).

Keywords


None

IDLmlTanHNormalizer::Restore


The IDLmlTanHNormalizer::Restore static method restores the normalizer from a file.

Syntax


Result = IDLmlTanHNormalizer.Restore(Filename)

Return Value


A reference to the object instance restored from the file.

Arguments


Filename

Specify the name of the file to restore.

Keywords


None

IDLmlTanHNormalizer::Save


The IDLmlTanHNormalizer::Save method saves the normalizer to a file.

Syntax


Obj->[IDLmlTanHNormalizer::]Save, Filename

Arguments


Filename

Specify the name of the file to save.

Keywords


None

IDLmlTanHNormalizer::Unnormalize


The IDLmlTanHNormalizer::Unnormalize method returns normalized data to its original values.

Syntax


This method can be invoked as a procedure or a function. The procedure syntax is faster and more efficient, but it overwrites the Data argument. The function syntax is less efficient, but leaves the input argument untouched.

As a procedure:

Obj->[IDLmlTanHNormalizer::]Unnormalize, Data

As a function:

Result = Obj->[IDLmlTanHNormalizer::]Unnormalize(Data)

Return Value


If called as a procedure, the unnormalized array will overwrite the input Data argument. If called as a function, the return value will be an array with the same dimensions as Data, which contains the unnormalized result.

Arguments


Data

Specify a scalar, 1D array or 2D array of data to normalize.

If the data is 2D, the array must be of size n x m, where n is the number of attributes and m is the number of examples. If the data is 1D, the array must be of size n (number of attributes).

Keywords


None

Version History


8.7.1

Introduced

See Also


IDLmlLinearNormalizer, IDLmlRangeNormalizer, IDLmlUnitNormalizer, IDLmlVarianceNormalizer



© 2019 Harris Geospatial Solutions, Inc. |  Legal
My Account    |    Store    |    Contact Us