BTLLasso

Performs 'BTLLasso' as described by Schauberger and Tutz (2019) doi:10.18637/jss.v088.i09 and Schauberger and Tutz (2017) doi:10.1177/1471082X17693086. BTLLasso is a method to include different types of variables in paired comparison models and, therefore, to allow for heterogeneity between subjects. Variables can be subject-specific, object-specific and subject-object-specific and can have an influence on the attractiveness/strength of the objects. Suitable L1 penalty terms are used to cluster certain effects and to reduce the complexity of the models.