Which is a good choice LDA or PCA for feature reduction in the supervised learning model?

Question

PCA -> Unsupervised Model or use for supervise learning too LDA -> supervise Model Both used for the feature reduction.

Which is batter LDA or PCA for supervising learning feature reduction and why?

Data-set: It is very famous data-set of wine to find out the customer category.

score 1 · Accepted Answer · answered May 19 '18 at 10:30

1

If you have labels, a supervised approach will usually be much better than an unsupervised approach.

At least if the labels suit your problem.

If you do not have labels, then you can't use LDA.

answered May 19 '18 at 10:30

Pca is used for both supervise or unsupervised approach. It mean's that if we are in the domain of supervise learning then must go for the LDA. – Sohaib Aslam May 19 '18 at 14:55
1

PCA is unsupervised. Even when used on labeled data - it's not using the labels. – Has QUIT--Anony-Mousse May 19 '18 at 19:27

1 Answers1