I have divided my data into a training set and a test set. I have then applied a logistic transformation to the variables in the training set and have used pam to assign the observations to one of four clusters. My question is How do I score the test observations now that I have the training set with clusters? Thank you. -- __________________________ *Barry E. King, Ph.D.* __________________________ [[alternative HTML version deleted]]
Aaditya Nanduri
2014-May-27 15:23 UTC
[R] Scoring a test set after clustering a training set
Hi Barry, I don't think pam has a way to predict for test observations. However, this questions was asked and answered earlier. The approach is to take the medoids generated in pam and use them in a K-nn algorithm as the training data with K = 1. This way, you can classify your testing data by identifying the closest medoid (and therefore, the cluster it belongs to). Please see the conversation here: https://stat.ethz.ch/pipermail/r-help/2004-January/044511.html Regards, Aaditya On Mon, May 26, 2014 at 12:50 PM, Barry King <barry.king@qlx.com> wrote:> I have divided my data into a training set and a test set. > I have then applied a logistic transformation to the variables > in the training set and have used pam to assign the observations > to one of four clusters. > > My question is How do I score the test observations now that I have the > training set with clusters? > > Thank you. > > -- > __________________________ > *Barry E. King, Ph.D.* > __________________________ > > [[alternative HTML version deleted]] > > ______________________________________________ > R-help@r-project.org mailing list > https://stat.ethz.ch/mailman/listinfo/r-help > PLEASE do read the posting guide > http://www.R-project.org/posting-guide.html > and provide commented, minimal, self-contained, reproducible code. >[[alternative HTML version deleted]]