We apply multidimensional item response
theory models (MIRT) to analyse multi-category
purchase decisions. We further compare
their performance to benchmark models by
means of topic models. Estimation is based
on two types of data sets. One contains only
binary the other polytomous purchase decisions.
We show that MIRT are superior
w. r. t. our chosen benchmark models. In particular,
MIRT are able to reveal intuitive latent
traits that can be interpreted as characteristics
of households relevant for multi-category
purchase decisions. With the help of latent
traits marketers are able to predict future purchase
behaviour for various types of households.
These information may guide shop
managers for cross selling activities and product
recommendations.
| Identifer | oai:union.ndltd.org:VIENNA/oai:epub.wu-wien.ac.at:6538 |
| Date | January 2017 |
| Creators | Schröder, Nadine |
| Publisher | Vahlen |
| Source Sets | Wirtschaftsuniversität Wien |
| Language | English |
| Detected Language | English |
| Type | Article, PeerReviewed |
| Format | application/pdf |
| Rights | Creative Commons: Attribution-NonCommercial-NoDerivatives 4.0 International (CC BY-NC-ND 4.0) |
| Relation | https://doi.org/10.15358/0344-1369-2017-2-27, https://elibrary.vahlen.de/, https://elibrary.vahlen.de/zeitschrift/0344-1369, http://epub.wu.ac.at/6538/ |
Page generated in 0.0019 seconds