Bias reduction in exponential family nonlinear models

By Ioannis Kosmidis, David Firth. 2009

213 citations recorded by the source corpus.

View Paper: https://www.semanticscholar.org/paper/53c39b1000c42d9291b2e26ee5b7b46e146ad96a

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Statistics and Probability

Citations in the corpus

Stable isotopes confirm a foraging dichotomy in juvenile loggerhead sea turtles (2010, methodology): “Regression analyses were performed in R using the BRGLM package (Bias Reduced Generalized Linear Model; Kosmidis and Firth, 2008) with a binomial family and logit link function; the BRGLM function penalizes themaximum likelihood estimate for cases of perfect separation in the data (i.”
Multinomial logit bias reduction via the Poisson log-linear model (2011, result): “This might at first sight appear to complicate what is intended to be a simplifying computational device; however, the general results of Kosmidis & Firth (2009) apply, and yield a useful representation of the adjusted score vector which in turn suggests a simple iterative algorithm.”
Multilevel Analysis with Few Clusters: Improving Likelihood-Based Methods to Provide Unbiased Estimates and Accurate Inference (2020, methodology): “That is, coefficient estimates tend to be systematically and substantially larger (in absolute size) than the corresponding true values when the sample size is small; see Firth (1993); Kosmidis and Firth (2009); Zorn (2005).”
Maximum penalized likelihood estimation for skew-normal and skew-t distributions (2012, methodology): “This is the route taken Firth (1993) where the requirement of bias reduction is adopted at the onset of the construction; see also further work by Kosmidis & Firth (2009).”
Maximum Likelihood and Firth Logistic Regression of the Pedestrian Route Choice (2017, methodology): “To support their argument, Heinze and Schemper (2002) used two large-scale empirical studies in the binary logistic regression settings and it was validated later by Kosmidis (2007).”
Median bias reduction of maximum likelihood estimates (2016, unknown): “Indeed, one major advantage of the approach in Firth (1993) and Kosmidis & Firth (2009) is that the modified estimating equation does not depend explicitly on the maximum likelihood estimate.”