Le procedure di analisi multicriterio a supporto della stima immobiliare: il dominance-based rough set approach e l'UTA

Since the beginning of XX century appraisers have been trying to introduce in the trade practice quantitative approaches developed from mathematical and statistical fields. The goal was to introduce a formal, objective and clear approach to real estate appraisal, over and above the usual heuristic approaches. The academic researches on quantitative methods spread over the past fifty years in particular on two main fields: on one hand an in-deep examination of the market value’s forming mechanism, on the other hand setting-up robust and clear estimation models. Both these research fields have produced a great amount of works that has helped to steady the appraisal’s scientific profile. Nevertheless rebound in trade practice appraiser’s activities seems small, as to hedge the application only in few circumstance as mass appraisal. Reason behind are known and essentially having reference to lack of availability and quality data, and the need to test the data’s strength, pliability, and transparency. These reasons together with a need for procedure able to taking into account also appraisers knowledge and judgment suggest the application of multiplecriteria methods for real estate appraisal. The aim of this work is to propose the application of two methods developed in multicriteria decision analysis (MCDA) as real estate appraisal tool: the dominance-based rough set approach (DRSA) and the UTA method. Classical Rough Set Approach (CRSA) has been recently proposed as an automated valuation methodology for mass appraisal. In CRSA market data are organized in value classes and the procedure define a set of “if … then …” rules based on a “crisp” indiscernibility relations between the classes. Then rules are used to classify the properties in the appropriate value classes. To overcome the limits of the “crisp” rules a value tolerance relation (VRA) has been introduced to improve flexibility in rules generation. The procedure proposed in this work is based on Dominance-based Rough Set Approach (DRSA) an extension of the CRSA suitable for valuation problems where the order properties of real property characteristics (surface, age, etc.) and the sales price is properly taken into account. In DRSA indiscernibility principle is substituted by dominance principle and this improves the amount of information used by the model in defining rules. In particular DRSA does not need to organize market data in value classes because it permits to estimate a value interval for a specific property in terms of “at least” and “at most” values. UTA method is a procedure able to asses a set of utility functions, consistent with the decision-maker’s a priori preferences. The set of utility functions is assessed using an ordinal regression method and linear programming. Linear programming is used to optimally adjust additive non-linear utility functions. This method could be proficiently used as adjustments factors estimate in the Adjustment Grid Method (AGM). An application of DRSA and UTA on a database of property gather in the city of Trieste (Italy) proves their usefulness for real estate appraisal.

Negli ultimi vent'anni si è registrata una particolare attenzione da parte della letteratura estimativa verso la studio e la ricerca di nuovi approcci alla stima immobiliare. Tale attenzione è, in parte, da attribuirsi alle nuove richieste di un mercato immobiliare in evoluzione che richiede al perito estimatore una sempre maggiore attenzione ai processi che permettono di giungere ad un giudizio estimativo trasparente e verificabile. Oltre ad affinare i consolidati approcci di carattere statistico-econometrico la disciplina ha cominciato ad esplorare le metodologie e modelli sviluppati in altri campi di ricerca. In particolare è stato rilevato come approcci propri dell’analisi multicriteriale possano essere applicabili con profitto al campo estimativo. Il lavoro di ricerca proposto sviluppa e approfondire l’applicazione di due metodi sviluppati nel campo dell'MCDA, recentemente applicati anche al campo estimativo, la teoria Rough Set e l'UTA. La ricerca ha potuto evidenziare l'utilità di tali metodi come supporto all'attività estimativa di predisposizione del giudizio di stima su base sintetico-comparativa.