Commercial Properties Prices Appraisal: Alternative Approach Based on Neural Networks

Abstract

There are different approaches to obtain the market price of a dwelling. Several agents, such as buyers and sellers, or local or tax authorities need to estimate the value of properties. Econometric hedonic models encounter several theoretical and practical difficulties when applied to the real estate market, such as downward biases in the estimation of hedonic prices, subjective decisions in the measurement process of categorical attributes, frontier problems related to an imperfect information framework, uniequational specification. Many of these are linked to the parametric approach. A lot of papers have been produced in the academic literature for such purposes, but, these are, almost always, oriented to estimate hedonic prices of residential properties, such as houses or apartments. Here these methodologies are used in the field of estimate market price of commercial premises, using Artificial Intelligence (AI) techniques. A case study is developed in Cordova –a medium city in the South of Spain-. Artificial Neural Networks (ANN) are an attractive alternative to the traditional hedonic modelling approaches, as they are better adapted to non-linearities of causal relationships and they also produce smaller valuation errors. It is also possible, from the ANN model, to obtain implicit prices associated to the main attributes that can explain the variability of the market price of commercial properties.