Dimension thresholds for log-concavity of normalized mixed moments
Loading original first page…
Abstract
For a nonnegative coordinatewise nondecreasing integrable weight on a finite d-dimensional box, we prove joint log-concavity of its mixed moments divided by ∏_iΓ(α_i+1) whenever α_i≥ d/4, d≥2. The weight need not be log-concave, independent across coordinates, or positively associated. For finite product grids with ordered log-concave coordinate values and a weight nondecreasing in the coordinate indices, the corresponding sufficient orthant is α_i≥ d/4+4. Increasing upper-corner densities and finite-grid weights give counterexamples at common exponent d/4-O(1). Thus the universal dimension thresholds in the two classes both have optimal leading constant 1/4. The proof combines conditional logarithmic-deficit bounds with a signed-block covariance estimate and elementary trigamma inequalities. Explicit rational finite-grid and continuous counterexamples accompany the asymptotic construction. The same method gives spatial product-power smoothing of arbitrary decreasing densities and arrays at powers d/4+4 and d/4+7, respectively, again with optimal leading coefficient 1/4. A classical matroid log-concavity input yields a dimension-free binary endpoint, whereas unweighted hereditary binary supports already require dimension-linear powers.
Read the full paper in the original PDF.
Edition files
- Paper.pdf136.7 KiB · pdf
SHA-256
a19e9c3cb0ec5119071a87e6174d22e586f0605329a87af784b319247b6a1c32 - Publication_Package.zip158.4 KiB · package
SHA-256
02ac45e642b129faf2e4b7ad3d63186e46540718d960927bcbe1011ee616c12d
Available editions
Research details
- Subjects
- Mathematics · Probability and Statistics · Classical and Real Analysis
- Keywords
- Classical and Real Analysis, Probability and Statistics
- Manuscript date
- 2026-10-24
- Reader reuse
- No additional permission granted
Scope
Joint log-concavity of Gamma-normalized mixed moments and spatial power smoothing for arbitrary monotone inputs, with optimal leading dimension coefficient 1/4 and distinct binary-structure thresholds.
Disclosures
AI-assisted tools supported research exploration, literature review, manuscript preparation and internal verification. The author retains responsibility for the work. Internal AI-assisted and computational checks are distinct from external human peer review.
Reproduction
15-page manuscript and 6 package members. Compile main.tex with compatible LaTeX or Tectonic. Run python -B verify_moments.py and python -O -B verify_moments.py using the Python 3 standard library; both print the same exact rational witness and polynomial-certificate results without reading external data or writing files.