To date, the approaches to modelling housing preferences, have mostly been quantitative. This working paper explores an alternative approach, namely that of decision tables that focus on conditions and states leading to particular actions or decisions. The approach is illustrated by modelling people's preferences in choosing a new dwelling.
The decision tables comprise sets of decision rules derived from secondary analyses of an existing data set, using a CHAID-based algorithm. Decision tables make for an attractive approach in the context of rule-based (parts of) micro-simulation models of consumer choice behaviour. The approach is illustrated by modelling people's preferences in choosing a new dwelling. In addition, the flexibility of the approach is used to demonstrate how the characteristic of 'housing affordability' is related to both the desired dwelling and the potential buyer or tenant. Rules are extracted from data on stated preferences, collected in the Netherlands in 2009.
PBL Working paper 16