Optimize Selection before naive bayes modelling

norita
norita New Altair Community Member
edited November 5 in Community Q&A
Heello

In the tutorial process the operator optimize selection contains svm (support vector machine). I was wondering if I can use it alike for feature selection. I will use afterwards naive bayes. Should I have any concerns that this different model methods doesn't fit together?

Best

Nora

Tagged:

Best Answer

  • earmijo
    earmijo New Altair Community Member
    Answer ✓
    Hi Norita: Yes, you could use other algorithms. However, the selection of the optimal set of variable may not be the same. There are two ways of doing variable selection: wrappers and filters. Filters give you the same answer for all algorithms. You select the variable before you choose the algorithm. With Wrappers (and Optimize Selection is the operator to use in RM) the selection is dependent on the algorithm you choose. 

Answers

  • earmijo
    earmijo New Altair Community Member
    Answer ✓
    Hi Norita: Yes, you could use other algorithms. However, the selection of the optimal set of variable may not be the same. There are two ways of doing variable selection: wrappers and filters. Filters give you the same answer for all algorithms. You select the variable before you choose the algorithm. With Wrappers (and Optimize Selection is the operator to use in RM) the selection is dependent on the algorithm you choose.