Generalizing Pearson's coefficient to determine properties of embedded manifold

41 Views Asked by At

I have the following dilemma:

We know that for random vectors we have Pearson's coefficient of skewness. I think you all agree that in some sense it measures the shape properties of the distribution.

Now let's generalize this: we have a set of vectors (dependent or independent). Could we generalize the Pearson's R in a way to determine some properties of the embedded manifold such as shape and dimension?