Error on linear discriminant analysis when applied Optical Recognition of Handwritten Digits Data Set

[sleung]

Data Set:
https://archive.ics.uci.edu/ml/machine-learning-databases/optdigits/optdigits.tra

I got an error on this data set when using:

04_Performing_a_Linear_Discriminant_Analysis [Workflow] — NodePit

Or using Linear Discriminant Analysis Compute node

Error is : cannot invert the inter-class scatter matrix as it is singular. Most Likely, two input columns are linearly dependent, ie. differ only by a constant factor.

[janina]

Hallo @sleung,

as the error message states, you probably have some linearly dependent input columns. You can remove those input columns using the Linear Correlation node and the Linear Correlation Filter node.
I hope that helps. Let me know if you have further questions.

Kind regards,
Janina