Regression trees for hospitality data analysis

Abstract

Purpose The purpose of this note is to describe the concept of regression trees (RTs) for hospitality data analysis. Design/methodology/approach RT is an effective non-parametric predicting modelling approach that would free researchers from the need to force a certain functional form. The method does not require normalization or scaling of data. Findings The authors illustrate how RTs can be used to find a model that would result in the best prediction. Research limitations/implications A common challenge facing hospitality researchers is to estimate a regression model with the correct specification. RTs can help researchers identify the best explanatory model for prediction. Originality/value This paper describes the concept of RTs for the modelling of hospitality data.