Faq build a classification model from class set other than the first

From Eigenvector Research Documentation Wiki
Revision as of 12:58, 21 November 2018 by imported>Lyle
Jump to navigation Jump to search

Issue:

How do I build a classification model from a class set other than the first?

Possible Solutions:

As of version 6.5 of PLS_Toolbox and Solo, the class set to model can be selected from either the Class Groups interface (in PLSDA) or in the method settings

  • (See the Edit > Options > Method Options menu or the Method Options toolbar button).


In earlier versions of the software, the following instructions indicate how to use other class sets:

Although DataSet objects allow you to use multiple class sets to identify different groupings of your samples, most classification methods (e.g. SIMCA, PLSDA, SVMDA, KNN) only allow you to use the first class set as the basis for the model.

To use a class set other than the first as the basis for a classification model, you need to copy the class set you want into the first set.

  1. Start up the DataSet editor (double-click the "X" block icon) go the Row Labels tab. On that tab, you will find the class sets listed in a pull-down menu just below the "Class" column header.
  2. Select the class set you want to build your model from in the pull-down menu.
  3. Right-click the "Class" column header and select "copy".
  4. Switch the class set pull-down menu back to the first set (top of the list).
  5. Right-click the "Class" column header again and select "paste"

Ability to use a class set other than the first is expected to be added in a later version of the software.