All Tools Bookmark


Facebook Share Twitter Share

BTLLasso: Modelling Heterogeneity in Paired Comparison Data

Performs 'BTLLasso', a method to model heterogeneity in paired comparison data. Subject-specific covariates are allowed to have an influence on the attractivity/strength of the objects. An L1 penalty on the pairwise differences between the object-specific parameters allows for both clustering of objects with regard to covariates and elimination of irrelevant covariates.