A COMPARATIVE ANALYSIS OF ARTIFICIAL NEURAL NETWORKS AND HEDONIC PRICING IN PREDICTING LAGOS PROPERTY PRICES

The search for techniques that can achieve better valuation estimates is a global issue, while much research has been done in the advanced countries, only a few researches are known to have been done in the area of adopting the advanced techniques in predicting property prices in Nigeria. The current paper compares the performance of the hedonic pricing model (HPM) and artificial neural networks (ANN) models in predicting the prices of residential properties in Lagos metropolis, Nigeria. Residential property prices and data for the variables affecting property prices were obtained from the databases of 53 firms of practicing Estate Surveyors and Valuers in the study area. A total of 3,079 datasets, encompassing property, neighborhood, and environmental-based features were gathered and employed in the research along with 19 explanatory factors. The entire dataset was split into training and testing at a ratio of 80% and 20% respectively to assess the prediction ability of the ANN and HPM. For both models, the performance evaluation metrics of R-squares, mean absolute error (MAE), mean absolute percentage error (MAPE), and root mean square error (RMSE) were computed and then compared. ANN model outperformed HPM model in predicting residential property prices in the Lagos metropolitan residential property market. The outcome of the research provides decision inputs for policymakers, investors in real estate, real estate professionals, and other stakeholders