One of the main issues of traditional predictive models on real estate is that housing prices are largely estimated based on the hedonic pricing theory, analyzing properties’ attributes with methods such as regressions and neural networks. None of these techniques quantifies the -similarity- among properties, a major component on real estate markets.
This model proposes a network science approach to incorporate the effects of similarity on price estimators. The amount of price-changing attributes among properties creates an economic network of similar assets, employed to estimate prices.
The network-based valuation model could potentially increment traditional hedonic pricing analysis and establish an enlightening perspective about real estate economic-network dynamics.
More information coming soon!