Abstract | ||
---|---|---|
We study the problem of approximating the recovery of a probability distribution on the unit interval from its first k moments. As main result we obtain an upper bound on the L-1 reconstruction error under the regularity assumption that the log-density function has square-integrable derivatives up to some natural order r > 1. Our bound is of order O(k(-r)). A comparative study relates our findings to alternative conditions on the distributions. (C) 2021 Elsevier Inc. All rights reserved. |
Year | DOI | Venue |
---|---|---|
2021 | 10.1016/j.amc.2021.126057 | APPLIED MATHEMATICS AND COMPUTATION |
Keywords | DocType | Volume |
Truncated Hausdorff moment problem, Moment-based distribution approximation, Total variation distance, Maximum entropy | Journal | 400 |
ISSN | Citations | PageRank |
0096-3003 | 0 | 0.34 |
References | Authors | |
0 | 2 |
Name | Order | Citations | PageRank |
---|---|---|---|
Werner Zellinger | 1 | 32 | 4.27 |
Bernhard Alois Moser | 2 | 0 | 0.34 |