Someone said that I need to define "intrinsic dimensionality" in my
question.
Basically, this is the topological dimension of a set of points that are
represented in a space of possibly higher dimension. For example, your data
might describe a corkscrew line through a 30-dimensional space (you measured
30 variables), so the intrinsic dimension of your set of points is 1.
A reference is Trunk, 1976, Statistical Estimation of the Intrinsic
Dimensionality of a Noisy Signal Collection, IEEE Transactions on Computers,
Vol c-25, No 2, pp 165-171. There are, however, many other algorithms.
--
View this message in context:
http://r.789695.n4.nabble.com/estimate-intrinsic-dimension-tp2956401p2956607.html
Sent from the R help mailing list archive at Nabble.com.