Optimal linear subdivision rules for noisy data

Subdivision schemes are iterative processes that recursively refine data by applying subdivision rules. This paper introduces linear subdivision rules tailored to handle noisy data. A key innovation lies in determining the rule coefficients by solving an optimization problem aimed at minimizing the noise variance. The study addresses the general case, allowing for noise correlation among data with a non-uniform distribution. In fact, we show that the subdivision rules, proposed in [S. L\'opez-Ure\~na and D. F. Y\'a\~nez, J. Sci. Comput., 100(1) (2024)], are optimal for uncorrelated noise with non-uniform variance. Numerical experiments are provided to demonstrate the effectiveness of these optimal rules compared to other subdivision rules designed for noisy data.