Model comparison with composite likelihood information criteria

Ng, Chi Tim; Joe, Harry

doi:10.3150/13-bej539

Cited by 15 publications

(10 citation statements)

References 15 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…18 Because each network could have more than one possible pedigree, we evaluate all pairs of possible pedigrees identified by PRIMUS for each network and identify the pair of pedigrees that minimizes the CL-AIC of the two networks. For a given pair of pedigrees Net 1i and Net 2j , the CL-AIC is calculated according to Equation 4:…”

Section: Padre Algorithmmentioning

confidence: 99%

PADRE: Pedigree-Aware Distant-Relationship Estimation

Staples

Witherspoon

Jorde

et al. 2016

The American Journal of Human Genetics

View full text Add to dashboard Cite

Accurate estimation of shared ancestry is an important component of many genetic studies; current prediction tools accurately estimate pairwise genetic relationships up to the ninth degree. Pedigree-aware distant-relationship estimation (PADRE) combines relationship likelihoods generated by estimation of recent shared ancestry (ERSA) with likelihoods from family networks reconstructed by pedigree reconstruction and identification of a maximum unrelated set (PRIMUS), improving the power to detect distant relationships between pedigrees. Using PADRE, we estimated relationships from simulated pedigrees and three extended pedigrees, correctly predicting 20% more fourth- through ninth-degree simulated relationships than when using ERSA alone. By leveraging pedigree information, PADRE can even identify genealogical relationships between individuals who are genetically unrelated. For example, although 95% of 13(th)-degree relatives are genetically unrelated, in simulations, PADRE correctly predicted 50% of 13(th)-degree relationships to within one degree of relatedness. The improvement in prediction accuracy was consistent between simulated and actual pedigrees. We also applied PADRE to the HapMap3 CEU samples and report new cryptic relationships and validation of previously described relationships between families. PADRE greatly expands the range of relationships that can be estimated by using genetic data in pedigrees.

show abstract

Section: Padre Algorithmmentioning

confidence: 99%

PADRE: Pedigree-Aware Distant-Relationship Estimation

Staples

Witherspoon

Jorde

et al. 2016

The American Journal of Human Genetics

View full text Add to dashboard Cite

show abstract

“…where n is the number of observations of the data series x. Note that the various models we consider are generally non-nested, whereas most research on model selection using the composite likelihood approach has considered nested model (Ng and Joe, 2014). An analysis of the properties of CLAIC and CLBIC in the non-nested case in the spirit of, e.g., Vuong (1989) would be very valuable but is beyond the scope of the present article.…”

Section: Information Criteria For Model Selectionmentioning

confidence: 99%

Inference and forecasting for continuous-time integer-valued trawl processes

Bennedsen¹,

Lunde²,

Shephard³

et al. 2021

Preprint

View full text Add to dashboard Cite

This paper develops likelihood-based methods for estimation, inference, model selection, and forecasting of continuous-time integer-valued trawl processes. The full likelihood of integervalued trawl processes is, in general, highly intractable, motivating the use of composite likelihood methods, where we consider the pairwise likelihood in lieu of the full likelihood. Maximizing the pairwise likelihood of the data yields an estimator of the parameter vector of the model, and we prove consistency and asymptotic normality of this estimator. The same methods allow us to develop probabilistic forecasting methods, which can be used to construct the predictive distribution of integer-valued time series. In a simulation study, we document good finite sample performance of the likelihood-based estimator and the associated model selection procedure. Lastly, the methods are illustrated in an application to modelling and forecasting financial bid-ask spread data, where we find that it is beneficial to carefully model both the marginal distribution and the autocorrelation structure of the data. We argue that integer-valued trawl processes are especially well-suited in such situations.

show abstract

“…This definition focuses on the marginal distributions, but also imposes that the compared CML models are composed of marginals with the same dimension. Ng and Joe (2014) address some of the differences between the CLAIC and normal AIC. To do so they focus on nested models with tractable likelihoods and make use of the theory of local alternatives.…”

Section: Information Criteriamentioning

confidence: 99%

Model selection and model averaging in MACML-estimated MNP models

Batram,

Bauer

2017

Preprint

View full text Add to dashboard Cite

This paper provides a review of model selection and model averaging methods for multinomial probit models estimated using the Maximum Approximate Composite Marginal Likelihood (MACML) approach. The proposed approaches are partitioned into test based methods (mostly derived from the likelihood ratio paradigm), methods based on information criteria and model averaging methods. Many of the approaches first have been derived for models estimated using maximum likelihood and later adapted to the composite marginal likelihood framework. In this paper all approaches are applied to the MACML approach for estimation. The investigation lists advantages and disadvantages of the various methods in terms of asymptotic properties as well as computational aspects. We find that likelihoodratio-type tests and information criteria have a spotty performance when applied to MACML models and instead propose the use of an empirical likelihood test. Furthermore, we show that model averaging is easily adaptable to Composite Marginal Likelihood (CML) estimation and has promising performance w.r.t to parameter recovery. Finally model averaging is applied to a real world example in order to demonstrate the feasibility of the method in real world sized problems.

show abstract

Model comparison with composite likelihood information criteria

Cited by 15 publications

References 15 publications

PADRE: Pedigree-Aware Distant-Relationship Estimation

PADRE: Pedigree-Aware Distant-Relationship Estimation

Inference and forecasting for continuous-time integer-valued trawl processes

Model selection and model averaging in MACML-estimated MNP models

Contact Info

Product

Resources

About