Locally interpretable tree boosting: An application to house price prediction
Peer reviewed, Journal article
Accepted version
Permanent lenke
https://hdl.handle.net/11250/3104580Utgivelsesdato
2023Metadata
Vis full innførselSamlinger
- NTNU Handelshøyskolen [1645]
- Publikasjoner fra CRIStin - NTNU [38679]
Originalversjon
https://doi.org/10.1016/j.dss.2023.114106Sammendrag
We introduce Locally Interpretable Tree Boosting (LitBoost), a tree boosting model tailored to applications where the data comes from several heterogeneous yet known groups with a limited number of observations per group. LitBoost constraints the complexity of a Gradient Boosted Trees model in a way that allows us to express the final model as a set of local Generalized Additive Models, yielding significant interpretability benefits while still maintaining some of the predictive power of a Gradient Boosted Trees model. We use house price prediction as a motivating example and demonstrate the performance of LitBoost on a data set of observations from different city districts in Oslo (Norway). We also test the robustness of LitBoost in an extensive simulation study on a synthetic data set.