Proposes an adaptive hybrid estimator for common mean estimation under independent but non-identical symmetric unimodal distributions, with near-optimality guarantees even when only log n / n samples are low-noise.
Combining the inequalities, we conclude that min ˆµ max {Pi}⊆P (s1,s 2,p ) E[∥ˆµ− µ∥2]≥ s ( min ˆµ max µ Pµ Bin(∥ˆµ− µ∥2≥ s)− 2 exp(−c′logn) )
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
math.ST 1years
2019 1verdicts
UNVERDICTED 1representative citing papers
citing papers explorer
-
Estimating location parameters in entangled single-sample distributions
Proposes an adaptive hybrid estimator for common mean estimation under independent but non-identical symmetric unimodal distributions, with near-optimality guarantees even when only log n / n samples are low-noise.