Traditional event models underlying naive Bayes classifiers assume probability distributions that are not appropriate for binary data generated by human behaviour. In this work, we develop a new event model, based on a somewhat forgotten distribution created by Kenneth Ted Wallenius in 1963. We show that it achieves superior performance using less data on a collection of Facebook datasets, where the task is to predict personality traits, based on likes.
![](https://fosterprovost.com/wp-content/uploads/2019/05/Line-42.png)
![](https://fosterprovost.com/wp-content/uploads/2019/05/Line-43.png)
![](https://fosterprovost.com/wp-content/uploads/2019/05/Line-41.png)
![](https://fosterprovost.com/wp-content/uploads/2019/05/Line-41.png)
Wallenius Naïve Bayes
- Enric Junque de Fortuny
- David Martens
- Foster Provost
- Venue: Machine Learning
- 2018
- Type: Journal Article