I want to know how I could perfom a cluster analysis with multiple features. Let's consider, for example, that I have a set of features for each object (I have a lot of objects). And each of these features has a particular dimension (Feature A is a vector of n dimensions, Feature B is a vector of m dimensions, etc.)
How can I cluster all the different features of all the objects together?
The typical thing is to just flatten your list of feature vectors into a combined feature vector whose length is the sum of the component feature vector lengths.
EG in your example with two feature vectors A and B, you'd make a feature vector of length n + m. Feature vector A with n features gets positions 0 to n - 1. Feature vector B with m dimensions gets positions n to n + m - 1.
