Presents an algorithm for support size testing using O(n/(ε log n) log(1/ε)) samples that nearly matches the Ω(n/(ε log n)) lower bound and improves on histogram learning.
Instance optimal learning of discrete distributions
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.DS 1years
2024 1verdicts
UNVERDICTED 1representative citing papers
citing papers explorer
-
Testing Support Size More Efficiently Than Learning Histograms
Presents an algorithm for support size testing using O(n/(ε log n) log(1/ε)) samples that nearly matches the Ω(n/(ε log n)) lower bound and improves on histogram learning.