Rare event analysis for minimum Hellinger distance estimators via large deviation theory
Publikation: Bidrag til tidsskrift › Tidsskriftartikel › Forskning › fagfællebedømt
Standard
Rare event analysis for minimum Hellinger distance estimators via large deviation theory. / Vidyashankar, Anand N.; Collamore, Jeffrey F.
I: Entropy, Bind 23, Nr. 4, 386, 2021.Publikation: Bidrag til tidsskrift › Tidsskriftartikel › Forskning › fagfællebedømt
Harvard
APA
Vancouver
Author
Bibtex
}
RIS
TY - JOUR
T1 - Rare event analysis for minimum Hellinger distance estimators via large deviation theory
AU - Vidyashankar, Anand N.
AU - Collamore, Jeffrey F.
PY - 2021
Y1 - 2021
N2 - Hellinger distance has been widely used to derive objective functions that are alternatives to maximum likelihood methods. While the asymptotic distributions of these estimators have been well investigated, the probabilities of rare events induced by them are largely unknown. In this article, we analyze these rare event probabilities using large deviation theory under a potential model misspecification, in both one and higher dimensions. We show that these probabilities decay exponentially, characterizing their decay via a \rate function," which is expressed as a convex conjugate of a limiting cumulant generating function. In the analysis of the lower bound, in particular, certain geometric considerations arise which facilitate an explicit representation, also in the case when the limiting generating function is non-differentiable. Our analysis also involves the modulus of continuity propertiesof the affinity, which may be of independent interest.
AB - Hellinger distance has been widely used to derive objective functions that are alternatives to maximum likelihood methods. While the asymptotic distributions of these estimators have been well investigated, the probabilities of rare events induced by them are largely unknown. In this article, we analyze these rare event probabilities using large deviation theory under a potential model misspecification, in both one and higher dimensions. We show that these probabilities decay exponentially, characterizing their decay via a \rate function," which is expressed as a convex conjugate of a limiting cumulant generating function. In the analysis of the lower bound, in particular, certain geometric considerations arise which facilitate an explicit representation, also in the case when the limiting generating function is non-differentiable. Our analysis also involves the modulus of continuity propertiesof the affinity, which may be of independent interest.
U2 - 10.3390/e23040386
DO - 10.3390/e23040386
M3 - Journal article
C2 - 33805183
VL - 23
JO - Entropy
JF - Entropy
SN - 1099-4300
IS - 4
M1 - 386
ER -
ID: 257193238