Using Multidimensional Item Response Theory Models to Explain Multi-Category Purchases

We apply multidimensional item response models (MIRT) to analyse multi-category decisions. We further compare performance to benchmark models by of topic models. Estimation is based two types of data sets. One contains only the other polytomous purchase decisions. show that MIRT are superior . r. t. our chosen benchmark models. In particular, are able to reveal intuitive latent that can be interpreted as characteristics households relevant for multi-category decisions. With the help of latent marketers are able to predict future purchase for various types of households. information may guide shop for cross selling activities and product .