Data dimensionality reduction

Updated at:

This topic describes the Data Dimensionality Reduction component.

Description

The Data Dimensionality Reduction component reduces the dimensionality of online data using a model generated by the Principal Component Analysis (PCA) component.

Calculation logic

PCA is a statistical method for dimensionality reduction. It transforms original variables into a new set of uncorrelated composite variables. A smaller number of these composite variables can then be selected to represent most of the information from the original variables.

Parameters

IN port - Input parameters

Parameter name

Description

Required

Data type

Model application

Select the model type and a specific model. Then, configure the model's input data.

Yes

Feature variables: Integer or floating-point number

Note

If non-numeric data exists, an exception is thrown.

OUT port - Output parameters

Parameter

Description

Required

Output data type

OUT

The number of output parameters corresponds to the number of principal components that you configured during model training. The parameters are named `pca_i`, where `i` is an integer that increments from 1 to `n`, and `n` is the number of principal components.

For example, if the number of principal components is 3, the component outputs three parameters: `pca_1`, `pca_2`, and `pca_3`.

No

Floating-point number

Output quality codes

The output quality code is determined as follows:

  • If the quality codes of all input variables are greater than or equal to 192, the quality code of the output variable is 192.

  • If the quality code of any input variable is in the range [0, 192), the quality code of the output variable is 0.

  • If all input variables have a quality code of -1, or if some have a quality code of -1 while the others are greater than or equal to 192, the quality code of the output variable is -1.