Batch Comparison

The PCA Batch Comparison view allows for the comparison of batch data for PCA models. Variable contributions are displayed per field for selected batches, or the user has the option to view a SPE plot for each batch in the model and comparison data.  

Variable contributions

Variable contributions are displayed per batch for user selected fields.  It is therefore possible to select multiple batches and compare variable contributions for multiple fields for each batch.  To access and view variable contributions:

  1. From the Batch Comparison view, select the Variable contribution option.

  2. Select the batches from the list of batches you wish to view.

  3. Select the fields for which you wish to view variable contributions per batch

Variable contributions are now displayed for each of the selected fields per batch.  Each batch is assigned a unique color.

Distance to Model

Distance to Model calculations can also be displayed per batch for selected batches on the Batch Comparison view.  The Distance to Model is calculated and displayed for each selected batch in the comparison and model data.  To access and view Distance to Model calculations:

  1. From the Batch Comparison view, select the Distance to Model option.

  2. Select the batches from the list of batches for which you wish to view Distance to Model calculations. You can use the Shift or Control (Ctrl) keys for making your selections.

Distance to Model calculations per batch are displayed. The following statistics concerning the Distance to Model are displayed in a table:

  • Average

  • Standard Deviation

  • 1 Sigma = Average + Standard Deviation

  • 2 Sigma = 1 Sigma + Standard Deviation

  • 3 Sigma = 2 Sigma + Standard Deviation

A cutoff threshold can be set as a reference to highlight the batches that exceed the threshold. Those batches that exceed the cutoff threshold are colored in an alternate color. By default, the cutoff threshold is set to 1 Sigma. The cutoff threshold is used to determine if a batch is unacceptable or not. The cutoff threshold is also stored when the PCA model is exported for use in a blueprint. Comparisons and decisions can then be made against this threshold.

Distance to Model Distribution

The Distance to Model Distribution is a histogram of the Distance to Model bar chart. This view is a numerical representation of the Distance to Model plot. The bin heights in the Distance to Model plot are the X-axis values in this plot. The Y-axis values are the number of occurrences of a particular Distance to Model value. The cutoff threshold can also be set in this view - thus allowing the user the ability to see how many batches exceed the cutoff threshold.


Related topics:

  

CSense 2023- Last updated: June 24,2025