Function Modelling Of The Market And Assessing The Degree Of Similarity Between Real Properties - Dependent Or Independent Procedures In The Process Of Office Property Valuation

Referring to the developed and presented in previous publications (e.g. Barańska 2011) two-stage algorithm for real estate valuation, this article addresses the problem of the relationship between the two stages of the algorithm. An essential part of the first stage is the multi-dimensional function modelling of the real estate market. As a result of selecting the model best fitted to the market data, in which the dependent variable is always the price of a real property, a set of market attributes is obtained, which in this model are considered to be price-determining. In the second stage, from the collection of real estate which served as a database in the process of estimating model parameters, the selected objects are those which are most similar to the one subject to valuation and form the basis for predicting the final value of the property being valued. Assessing the degree of similarity between real properties can be carried out based on the full spectrum of real estate attributes that potentially affect their value and which it is possible to gather information about, or only on the basis of those attributes which were considered to be price-determining in function modelling. It can also be performed by various methods. This article has examined the effect of various approaches on the final value of the property obtained using the two-stage prediction. In order fulfill the study aim precisely as possible, the results of each calculation step of the algorithm have been investigated in detail. Each of them points to the independence of the two procedures.