Subject-specific Bradley–Terry–Luce models with implicit variable selection

Casalicchio, Giuseppe; Tutz, Gerhard; Schauberger, Gunther

doi:10.1177/1471082x15571817

Cited by 6 publications

(9 citation statements)

References 34 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…In this case, the subjects can make their preference decisions on more than two preference categories. The works of Tutz [46], Agresti [47], Dittrich et al [48], and Casalicchio et al [49] provide extensions in this sense. However, they are unnecessary here, because our goal is to retrieve the underlying relative worth of each indicator in a simple way.…”

Section: Bradley-terry Model and Bsc Integrationmentioning

confidence: 97%

A Bradley-Terry Model-Based Approach to Prioritize the Balance Scorecard Driving Factors: The Case Study of a Financial Software Factory

et al. 2020

View full text Add to dashboard Cite

The prioritization of factors has been widely studied applying different methods from the domain of the multiple-criteria decision-making, such as for example the Analytic Hierarchy Process method (AHP) based on decision-makers’ pairwise comparisons. Most of these methods are subjected to a complex analysis. The Bradley-Terry model is a probability model for paired evaluations. Although this model is usually known for its application to calculating probabilities, it can be also extended for ranking factors based on pairwise comparison. This application is much less used; however, this work shows that it can provide advantages, such as greater simplicity than traditional multiple-criteria decision methods in some contexts. This work presents a method for ranking the perspectives and indicators of a balance scorecard when the opinion of several decision-makers needs to be combined. The data come from an elicitation process, accounting for the number of times a factor is preferred to others by the decision-makers in a pairwise comparisons. No preference scale is used; the process just indicates the winner of the comparison. Then, the priority weights are derived from the Bradley-Terry model. The method is applied in a Financial Software Factory for demonstration and validation. The results are compared against the application of the AHP method for the same data, concluding that despite the simplifications made with the new approach, the results are very similar. The study contributes to the multiple-criteria decision-making domain by building an integrated framework, which can be used as a tool for scorecard prioritization.

show abstract

Section: Bradley-terry Model and Bsc Integrationmentioning

confidence: 97%

A Bradley-Terry Model-Based Approach to Prioritize the Balance Scorecard Driving Factors: The Case Study of a Financial Software Factory

et al. 2020

View full text Add to dashboard Cite

show abstract

“…Subject-specific covariates in paired comparisons were considered before by Turner and Firth (2012), Francis et al (2002) or Francis et al (2010). More recently, Casalicchio et al (2015) presented a boosting approach and Schauberger and Tutz (2017) a penalization approach based on an L 1 -penalty. Both are able to include explanatory variables and select the relevant ones.…”

Section: Subject-specific Covariatesmentioning

confidence: 99%

“…More recently, methods that are able to handle a larger number of explanatory variables have been proposed. Casalicchio, Tutz, and Schauberger (2015) presented a boosting approach and developed the corresponding R package ordBTL (Casalicchio 2014). The approach is restricted to subjectspecific covariates.…”

Section: Introductionmentioning

confidence: 99%

BTLLasso: A Common Framework and Software Package for the Inclusion and Selection of Covariates in Bradley-Terry Models

Schauberger¹,

Tutz²

2019

J. Stat. Soft.

Self Cite

View full text Add to dashboard Cite

In paired comparison models, the inclusion of covariates is a tool to account for the heterogeneity of preferences and to investigate which characteristics determine the preferences. Although methods for the selection of variables have been proposed no coherent framework that combines all possible types of covariates is available. There are three different types of covariates that can occur in paired comparisons, the covariates can either vary over the subjects, the objects or both the subjects and the objects of the paired comparisons. This paper gives an overview over all possible types of covariates in paired comparisons and introduces a general framework to include covariate effects into Bradley-Terry models. For each type of covariate, appropriate penalty terms that allow for sparser models and, therefore, easier interpretation are proposed. The whole framework is implemented in the R package BTLLasso. The main functionality and the visualization tools of the package are introduced and illustrated by real data sets.

show abstract

“…A boosting algorithm for differential item functioning in Rasch models was developed by Schauberger and Tutz [ 88 ] for the broader area of psychometrics, while Casalicchio et al focused on boosting subject-specific Bradley-Terry-Luce models [ 89 ].…”

Section: New Frontiers and Applicationsmentioning

confidence: 99%

An Update on Statistical Boosting in Biomedicine

Mayr

Hofner

Waldmann

et al. 2017

Computational and Mathematical Methods in Medicine

View full text Add to dashboard Cite

Statistical boosting algorithms have triggered a lot of research during the last decade. They combine a powerful machine learning approach with classical statistical modelling, offering various practical advantages like automated variable selection and implicit regularization of effect estimates. They are extremely flexible, as the underlying base-learners (regression functions defining the type of effect for the explanatory variables) can be combined with any kind of loss function (target function to be optimized, defining the type of regression setting). In this review article, we highlight the most recent methodological developments on statistical boosting regarding variable selection, functional regression, and advanced time-to-event modelling. Additionally, we provide a short overview on relevant applications of statistical boosting in biomedicine.

show abstract

Subject-specific Bradley–Terry–Luce models with implicit variable selection

Cited by 6 publications

References 34 publications

A Bradley-Terry Model-Based Approach to Prioritize the Balance Scorecard Driving Factors: The Case Study of a Financial Software Factory

A Bradley-Terry Model-Based Approach to Prioritize the Balance Scorecard Driving Factors: The Case Study of a Financial Software Factory

BTLLasso: A Common Framework and Software Package for the Inclusion and Selection of Covariates in Bradley-Terry Models

An Update on Statistical Boosting in Biomedicine

Contact Info

Product

Resources

About