deepnetts.data.norm

Class MaxScaler

java.lang.Object

deepnetts.data.norm.AbstractScaler

deepnetts.data.norm.MaxScaler

All Implemented Interfaces:

Serializable, javax.visrec.ml.data.preprocessing.Scaler<javax.visrec.ml.data.DataSet<MLDataItem>>

public final class MaxScaler
extends AbstractScaler

Performs max normalization, rescales data to corresponding max value in each column. Scales all values to interval [0, 1], by dividing columns with their corresponding maximum value. Performs normalization on both inputs and outputs.

See Also:

Serialized Form

Constructor Summary

Constructors

Constructor and Description
MaxScaler(javax.visrec.ml.data.DataSet<MLDataItem> dataSet) Creates a new instance of max normalizer initialized to max values in given data set.

Method Summary

Modifier and Type	Method and Description
void	apply(javax.visrec.ml.data.DataSet<MLDataItem> dataSet) Performs normalization on the given inputs.
void	deNormalizeInputs(Tensor inputs)
void	deNormalizeOutputs(Tensor outputs) De-normalize given output vector in-place.
Tensor	getMaxInputs()
Tensor	getMaxOutputs()
void	normalizeInput(Tensor input)
void	scaleInput(Tensor input) Normalize input of deployed model
void	setMaxInputs(Tensor maxInputs)
void	setMaxOutputs(Tensor maxOutputs)

Methods inherited from class java.lang.Object

equals, getClass, hashCode, notify, notifyAll, toString, wait, wait, wait

Constructor Detail

MaxScaler

public MaxScaler(javax.visrec.ml.data.DataSet<MLDataItem> dataSet)

Creates a new instance of max normalizer initialized to max values in given data set.

Parameters:

dataSet -

Method Detail

apply

public void apply(javax.visrec.ml.data.DataSet<MLDataItem> dataSet)

Performs normalization on the given inputs.

Parameters:

dataSet - data set to normalize

scaleInput

public void scaleInput(Tensor input)

Description copied from class: AbstractScaler

Normalize input of deployed model

Specified by:

scaleInput in class AbstractScaler

getMaxInputs

public Tensor getMaxInputs()

setMaxInputs

public void setMaxInputs(Tensor maxInputs)

getMaxOutputs

public Tensor getMaxOutputs()

setMaxOutputs

public void setMaxOutputs(Tensor maxOutputs)

deNormalizeOutputs

public void deNormalizeOutputs(Tensor outputs)

De-normalize given output vector in-place. Multiplies given vector with vector used for normalization, and stores these values in same memory location as input vector.

Parameters:

outputs -

deNormalizeInputs

public void deNormalizeInputs(Tensor inputs)

normalizeInput

public void normalizeInput(Tensor input)