Derivative and Savitzky-Golay cause incorrect predictions in 13.0.1 (Q758)

The information in this article applies to: 
SIMCA 13
Symptoms: 

Models built on Derivative or Savitzky-Golay filtered data, cause incorrect predictions for that model and all future predictions for models that include filtered data, whichever filters the models hold.

Workaround: 

Update to 13.0.2.

Status: 

This was fixed in 13.0.2.

More Information: 

When the used model is built on a dataset filtered with Derivative or Savitzky-Golay, the last variables in the predictionset hold nonsense values. These nonsense values thereafter cause incorrect predictions for that model and all future predictions for models built on filtered datasets. What this means is that predictions calculated after predictions for the Derivative/Savitzky-Golay model are incorrect for all filters. Only predictions from models built on original data are correct then.

Date modified: Oct 03, 2013
Article type: 
Problem