Synopsis
Determines the correlation between all attributes and can produce a weight vector based on correlations.
Description
This operator calculates the correlation matrix between all attributes of the input example set. Furthermore, attribute weights based on the correlations can be returned. This allows the de-selection of highly correlated attributes with the help of an AttributeWeightSelection operator. If no weights should be created, this operator produces simply a correlation matrix which up to now cannot be used by other operators but can be displayed to the user in the result tab.
Please note that this simple implementation performs a data scan for each attribute combination and might therefore take some time for non-memory example tables.
Input
- example set: expects: ExampleSet
Output
- example set:
- matrix:
- weights:
Parameters
- create weights: Indicates if attribute weights based on correlation should be calculated or if the complete matrix should be returned.
- normalize weights: Indicates if the attributes weights should be normalized.
- squared correlation: Indicates if the squared correlation should be calculated.