Abstract:The volume function V(t) of a compact set S\in R^d is just the Lebesgue measure of the set of points within a distance to S not larger than t. According to some classical results in geometric measure theory, the volume function turns out to be a polynomial, at least in a finite interval, under a quite intuitive, easy to interpret, sufficient condition (called ``positive reach'') which can be seen as an extension of the notion of convexity. However, many other simple sets, not fulfilling the positive reach condition, have also a polynomial volume function. To our knowledge, there is no general, simple geometric description of such sets. Still, the polynomial character of $V(t)$ has some relevant consequences since the polynomial coefficients carry some useful geometric information. In particular, the constant term is the volume of S and the first order coefficient is the boundary measure (in Minkowski's sense). This paper is focused on sets whose volume function is polynomial on some interval starting at zero, whose length (that we call ``polynomial reach'') might be unknown. Our main goal is to approximate such polynomial reach by statistical means, using only a large enough random sample of points inside S. The practical motivation is simple: when the value of the polynomial reach , or rather a lower bound for it, is approximately known, the polynomial coefficients can be estimated from the sample points by using standard methods in polynomial approximation. As a result, we get a quite general method to estimate the volume and boundary measure of the set, relying only on an inner sample of points and not requiring the use any smoothing parameter. This paper explores the theoretical and practical aspects of this idea.
What problem does this paper attempt to address?
### The problems the paper attempts to solve
The paper aims to solve the problem of how to estimate the polynomial reach of a compact set \( S \subset \mathbb{R}^d \) by statistical means. Specifically, the paper focuses on the set \( S \) where the volume function \( V(t) \) is a polynomial within certain intervals, and attempts to estimate this polynomial reach \( R \) through random sample points. This estimation method can be used to further estimate the volume \( \mu(S) \) and the boundary measure \( L(S) \) of the set \( S \).
### Background and motivation
1. **Volume function**:
- Given a compact set \( S \subset \mathbb{R}^d \) and \( r > 0 \), define the \( r \)-parallel set \( B(S, r) \) of \( S \) as the set of all points whose distance from \( S \) is no more than \( r \).
- The volume function \( V(r) \) is defined as \( V(r)=\mu(B(S, r)) \), where \( \mu \) represents the Lebesgue measure on \( \mathbb{R}^d \).
- When \( r = 0 \), \( V(0)=\mu(S) \).
- The boundary measure \( L(S) \) can be defined by the limit \( L(S)=\lim_{\epsilon \to 0^+}\frac{\mu(B(S, \epsilon)\setminus S)}{\epsilon} \).
2. **Polynomial volume property**:
- In many cases, the volume function \( V(r) \) is a polynomial within a certain interval \([0, R]\).
- For a set \( S \) with positive reach, its volume function is a polynomial within \([0, r]\), where \( r \) is the reach of \( S \).
- However, many simple sets may have their volume functions be polynomials within certain intervals even without positive reach.
3. **Statistical estimation**:
- Estimate the volume function \( V(r) \) through random sample points \( \aleph_n=\{X_1,\ldots,X_n\} \).
- Use the minimum distance method to estimate the coefficients of the volume function by fitting a polynomial.
- Estimating the polynomial reach \( R \) is the core problem in this paper, because it allows us to estimate the volume and the boundary measure of the set \( S \).
### Main contributions
1. **Uniform estimator**:
- Propose a uniform estimator \( \tilde{R} \) to estimate the polynomial reach \( R \).
- Prove that under certain conditions, \( \tilde{R} \) converges uniformly to \( R \).
2. **Lower bound algorithm**:
- Propose an algorithm to estimate the lower bound \( \hat{R} \) of the polynomial reach.
- This algorithm is safer with a finite sample size because it avoids errors that may be caused by over - estimation.
3. **Convergence rate**:
- Analyze the convergence rates of polynomial coefficient estimations, which depend on the dimension \( d \) and the sample size \( n \).
### Conclusion
Estimating the polynomial reach \( R \) by statistical means is an important problem. It not only helps to understand the geometric properties of sets, but also can be used to estimate the volume and the boundary measure of sets. The estimation methods proposed in this paper are of great significance in both theoretical and practical applications.