F_VN_InverseFeatureScalingExp

Apply the inverse of feature scaling to one or more sample(s). [fA, fB] should be used if the scaling has been performed using TCVN_FST1_MINMAX with a predetermined output range.. (expert function)

Syntax

Definition:

FUNCTION F_VN_InverseFeatureScalingExp : HRESULT
VAR_INPUT
    ipSamples        : ITcVnContainer;
    ipScales         : ITcVnContainer;
    ipInverseSamples : Reference To  ITcVnContainer;
    fA               : LREAL;
    fB               : LREAL;
    hrPrev           : HRESULT;
END_VAR

Inputs

Name	Type	Description
ipSamples	ITcVnContainer	Container holding one or more input sample(s) (ContainerType_Vector_REAL, ContainerType_Vector_LREAL, ContainerType_Vector_Vector_REAL, or ContainerType_Vector_Vector_LREAL)
ipScales	ITcVnContainer	Container with the scaling parameters of each feature (type ContainerType_Vector_REAL or ContainerType_Vector_LREAL).
ipInverseSamples	Reference To ITcVnContainer	Returns the inversed sample(s) using the scaling type and parameters of ipScales. It has the same type of ipSamples. If the same container ipSamples is used, the source data will be replaced.
fA	LREAL	represents the lower bound of the range for TCVN_FST1_MINMAX
fB	LREAL	represents the upper bound of the range for TCVN_FST1_MINMAX
hrPrev	HRESULT	HRESULT indicating the result of previous operations (If SUCCEEDED(hrPrev) equals false, no operation is executed.)

Return value

HRESULT

Further information

The function F_VN_InverseFeatureScalingExp is the expert variant of F_VN_InverseFeatureScaling. It contains additional parameters.

Parameter

Samples

One or more samples must be transferred as a container to ipSamples.

Scaling

A container with the scaling type and the corresponding scaling parameters must be transferred to ipScales. The container can be calculated using the function F_VN_GetFeatureScales or created using the structure described there.

Scaled samples

The inversely scaled samples are returned via the reference ipInverseSamples.

Defining the output range

If the scaling type TCVN_FST1_MINMAX is used, the value range of the output values can also be specified via fA and fB.

Application

For example, an inverse feature scaling where the scaling was previously calculated by F_VN_FeatureScalingExp with an output value range of 0.1 to 0.9 looks like this:

hr := F_VN_FeatureScalingExp(ipSamples, ipScales, ipSamplesScaled, 0.1, 0.9, hr);

hr := F_VN_InverseFeatureScalingExp(
    ipSamples       := ipSamplesScaled,
    ipScales        := ipScales,
    ipInverseSamples:= ipSamplesInvertedScaling,
    fA              := 0.1,
    fB              := 0.9,
    hrPrev          := hr);

Related functions

Required License

TC3 Vision Machine Learning

System Requirements

Development environment	Target platform	PLC libraries to include
TwinCAT V3.1.4024.54 or later	PC or CX (x64) with PL50, e.g. Intel 4-core Atom CPU	Tc3_Vision