Comparing PCA Data

This option allows the user to compare two different dataset's PCA models with each other.

1. Click on the Compare Data button.

2. Data source selection:

   1. Select the project data to use from All data, Brushed data or Non-brushed data. Note that only brushing that was performed before creating the model will be taken into consideration when using the Brushed or Non-brushed options. Once the PCA model has been created, you may brush in any of the trends without affecting the model.

   2. Select a different dataset to be used for comparison from the populated drop down list. These datasets are those imported or created during Data Preparation in Step 1.

   3. Apply limits: This will imposed the outer limits defined in the histograms on the new dataset fields for the PCA model comparison.  

   4. [Select] the dataset, then continue with mapping the new source’s inputs to those of the first data source’s PCA inputs. Click [Ok] when done.

3. Configure fields: define what to use as the batch index.

4. The two PCA models are then shown in the graph, the original in red and the comparison model in blue.
   
   Note: Compared data will be lost once the view is closed.

Clearing the comparison data

To remove the comparison data click on the Clear Comparison Data button.

Related topics:

• Constructing a PCA model

• PCA Bar Charts

• Constructing PLS model

• Batch Comparison