Handling invariant variables in TMPCA module

The TMPCA module is unstacking the multivariate process data used in MSPC analysis by time step. This can result in invariant variables being generated for certain time steps (e.g. if the value of one variable is the same for all batches at a time step).

In the current implementation, the Statistica TMPCA module cannot handle this situation and rejects to perform the analysis.

There are different ways to handle this situation (e.g. solving the PCA on the Covariance matrix instead of the Correlation matrix). It would be very important to solve this current limitation, in order to effectively use the TMPCA analysis in Statistica.

This limitation has also been discussed in support case 01866585.