carconf

Online configuration systems allow customers to actively participate in
the creation of products and become increasingly important in various
market sectors. Dabic and Hatzinger report a study on car configurators that aimed
at investigating the effects of certain person characteristics (such as gender)
on the configuration process. Subjects were asked to configure a car according
to their preferences. They could choose freely from several modules: such as
exterior and interior design, technical equipment, brand, price, and producing country.
The order of module choice was recorded as ranks. Since not all modules had to
be chosen the response format was partial rankings.

datasets

Generates design matrix for analysing real paired comparisons and derived paired comparison data (Likert type items/ratings or rankings) using a loglinear approach.  Fits loglinear Bradley-Terry model (LLBT) exploiting an eliminate feature.  Computes pattern models for paired comparisons, rankings, and ratings.  Some treatment of missing values (MCAR and MNAR). Fits latent class (mixture) models for paired comparison, rating and ranking patterns using a non-parametric ML approach.

Marco Johannes Maier

prefmod

Utilities to fit paired comparison models for preferences

carconf function

encoding

Dabic, M., Hatzinger, R. (2009). Zielgruppenadaequate Ablaeufe in
Konfigurationssystemen ?? eine empirische Studie im Automobilmarkt --
Partial Rankings. In R Hatzinger, R Dittrich, T Salzberger (eds.),
Praeferenzanalyse mit R: Anwendungen aus Marketing, Behavioural
Finance und Human Resource Management. Facultas.

carconf: Data (partial rankings): Car Configurator

Description

Usage

Arguments

encoding

source

Examples