SUMMARY
Many algorithms in machine learning and computational geometry require, as input, the intrinsic dimension of the manifold that supports the probability distribution of the data. This parameter is rarely known and therefore has to be estimated. We characterize the statistical difficulty of this problem by deriving upper and lower bounds on the minimax rate for estimating the dimension. First, we consider the problem of testing the hypothesis that the support of the data-generating probability distribution is a well-behaved manifold of intrinsic dimension d1d_1 versus the alternative that it is of dimension d2d_2, with d1<d2d_{1}. With an i.i.d. sample of size nn, we provide an upper bound on the probability of choosing the wrong dimension of O(n-(d2/d1-1-?)n)O\left( n^{-\left(d_{2}/d_{1}-1-\epsilon\right)n} \right), where ?\epsilon is an arbitrarily small positive number. The proof is based on bounding the length of the traveling salesman path through the data points. We also demonstrate a lower bound of O(n-(2d2-2d1+?)n)\Omega \left( n^{-(2d_{2}-2d_{1}+\epsilon)n} \right), by applying Le Cam's lemma with a specific set of d1d_{1}-dimensional probability distributions. We then extend these results to get minimax rates for estimating the dimension of well-behaved manifolds. We obtain an upper bound of order O(n-(1m-1-?)n)O \left( n^{-(\frac{1}{m-1}-\epsilon)n} \right) and a lower bound of order O(n-(2+?)n)\Omega \left( n^{-(2+\epsilon)n} \right), where mm is the embedding dimension.