PCA for classificatin problem.
Guru
dear all,
1. can we use PCA as dimens. reduction/feature selection technique for classification problem? If it is yes, how
to use both numeric feature and polynomial label in data set or PCA operator? (PCA operator accepts only numeric data)
2. If we cant use PCA for classification problem, how to do feature selection/dimensionality reduction for classification
problem using supervised algorithm.
thanks.
regds
thiru
1. can we use PCA as dimens. reduction/feature selection technique for classification problem? If it is yes, how
to use both numeric feature and polynomial label in data set or PCA operator? (PCA operator accepts only numeric data)
2. If we cant use PCA for classification problem, how to do feature selection/dimensionality reduction for classification
problem using supervised algorithm.
thanks.
regds
thiru
0
Answers
if the label is set correctly (Set Role operator), PCA will ignore the polynominal label. If you have other nominal attributes, you'll need an operator like Nominal to Numerical to recode it to numbers.
So you can easily use PCA in this situation.
Regards,
Balázs