Summary: |
The main purpose of this study is to understand how the price of housing is formed_in the new ( expo) and old town in Lisbon ( traditional neighborhoods). The study focuses on the prices of the apartments that were sold with the assistance of Estate Agents operating in Lisbon in last 3 years, and how the economic crisis can explain the market price. _._For the development of the study, we used two different methodologies: Hedonic_Pricing Methods (MPH) and Artificial Neural Networks (ANN). ANN are a less_traditional econometric technique from the field of Artificial Intelligence, but they_are strong competitors with the MPH. In the last two decades, the MPH has been_applied to the real estate market in Portugal but, till the present, no single study is_known with the ANN. To obtain the best hedonic model, numerous tests were_developed, aiming the validation of MPH and also the adequate selection of_variables that contribute most to the prices._The tests were performed with the statistical software SPSS, _The explanatory variables included in the final model for the price of an apartment_were: the floor area (m2), the garage and basement index, the comfort index, the_location index and two other variables resulting of interactions, one between the_year of sale and the condition of the apartment (whether it is new or used) and the_other between the preservation index and the condition of the apartment. With_these six explanatory variables, the MPH has achieved an accuracy quite_significant when compared with some previous studies. _ |