Using Bootstrap Likelihood Ratio in Finite Mixture Models

By Z. Feng, C. McCulloch. 1994

How researchers see this paper

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

Citations in the corpus

Residual Diagnostics for Growth Mixture Models (2005, background): “We note that in the sense of Redner (1981) and Feng and McCulloch (1996), (8) and (9) remain valid even under a relaxation of condition (I).”
A NONPARAMETRIC BAYESIAN APPROACH TO DETECT THE NUMBER OF REGIMES IN MARKOV SWITCHING MODELS (2002, methodology): “In classical terms, the problem is generally bypassed using the likelihood ratio test statistic, obtaining its distribution by simulation techniques; in particular, Feng and McCulloch([11]) provide the mathematical background to justify the use of bootstrap likelihood ratios.”
Finite mixture models and model-based clustering (2010, methodology): “Indeed, for the case of Gaussian mixtures with unequal variances, [39] recommended bootstrapping the LRT statistic over all other methods to avoid problems with regularity conditions.”
Identification and estimation of marginal effects in nonlinear panel models (2008, unknown): “We use the following binary choice model Yit = 1{Xitβ + αi + εit ≥ 0}, (17) 13”
Average and Quantile Effects in Nonseparable Panel Models (2009, unknown): “We use the following binary choice model Yit = 1{Xitβ + αi + εit ≥ 0}, (17) 13”
On a mixture autoregressive model (2000, background): “See Chen and Cheng (1997), Dacunha-Castelle and Gassiat (1997), Feng and McCulloch (1996) and Richardson and Green (1997).”
Editorial: recent developments in mixture models (2003, methodology): “In general, a parametric Bootstrap procedure can be used (McLachlan, 1992), and it was pointed out recently that this approach leads to valid statistical inference (Feng and McCulloch, 1996).”