Package jsat.datatransform

Interface Summary

Interface	Description
DataTransform	A pre-processing step may be desirable before training.
FastICA.NegEntropyFunc	The FastICA algorithm requires a function f(x) to be used iteratively in the algorithm, but only makes use of the first and second derivatives of the algorithm.
InPlaceInvertibleTransform	This interface behaves exactly as InPlaceTransform specifies, with the addition of an in-place "reverse" method that can be used to alter any given transformed data point back into an approximation of the original vector, without having to new vector object, but altering the one given.
InPlaceTransform	An In Place Transform is one that has the same number of categorical and numeric features as the input.
InvertibleTransform	A InvertibleTransform is one in which any given transformed vector can be inverse to recover an approximation of the original vector when using a transform that implements this interface.

Class Summary

Class	Description
AutoDeskewTransform	This transform applies a shifted Box-Cox transform for several fixed values of λ, and selects the one that provides the greatest reduction in the skewness of the distribution.
DataModelPipeline	A Data Model Pipeline combines several data transforms and a base Classifier or Regressor into a unified object for performing classification and Regression with.
DataTransformBase	This abstract class implements the Parameterized interface to ease the development of simple Data Transforms.
DataTransformProcess	Performing a transform on the whole data set before training a classifier can add bias to the results.
DenseSparceTransform	Dense sparce transform alters the vectors that store the numerical values.
FastICA	Provides an implementation of the FastICA algorithm for Independent Component Analysis (ICA).
Imputer	Imputes missing values in a dataset by finding reasonable default values.
InsertMissingValuesTransform	This transform mostly exists for testing code.
InverseOfTransform	Creates a new Transform object that simply uses the inverse of an InvertibleTransform as a regular transform.
JLTransform	The Johnson-Lindenstrauss (JL) Transform is a type of random projection down to a lower dimensional space.
LinearTransform	This class transforms all numerical values into a specified range by a linear scaling of all the data point values.
NominalToNumeric	This transform converts nominal feature values to numeric ones be adding a new numeric feature for each possible categorical value for each nominal feature.
NumericalToHistogram	This transform converts numerical features into categorical ones via a simple histogram.
PCA	Principle Component Analysis is a method that attempts to create a basis of the given space that maintains the variance in the data set while eliminating correlation of the variables.
PNormNormalization	PNormNormalization transformation performs normalizations of a vector x by one its p-norms where p is in (0, Infinity)
PolynomialTransform	A transform for applying a polynomial transformation on the data set.
RemoveAttributeTransform	This Data Transform allows the complete removal of specific features from the data set.
StandardizeTransform	This transform performs standardization of the data, which makes each column have a mean of zero and a variance of one.
UnitVarianceTransform	Creates a transform to alter data points so that each attribute has a standard deviation of 1, which means a variance of 1.
WhitenedPCA	An extension of PCA that attempts to capture the variance, and make the variables in the output space independent from each-other.
WhitenedZCA	An extension of WhitenedPCA, is the Whitened Zero Component Analysis.
ZeroMeanTransform	A transformation to shift all numeric variables so that their mean is zero

Enum Summary

Enum	Description
FastICA.DefaultNegEntropyFunc	A set of default negative entropy functions as specified in the original FastICA paper
Imputer.NumericImputionMode
JLTransform.TransformMode	Determines which distribution to construct the transform matrix from