MTシステムによる不動産価格の予測(3)

－単位空間が1データの場合のシュミットの直交展開を使用しない価格予測－

抄録

Real-estate appraisal so far has expressly considered the relatiqnship between real-estate price and greenery on the real-estate premises or in the surrounding region. The quality of the real estate setting, however, is a recognized pricing factor in the real-estate appraisal standard, and greenery undoubtedly has an effect on the economic valuation of real estate, because it constitutes part or all of the setting. In an effort to understand the influence of greenery on the economic value of real estate, regions in the suburbs of southwestern Tokyo were selected for appraisal and the Mahalanobis-Taguchi system was applied to the forecasting of their real estate prices. The analysis in this study was based on signal patterning, using a single-data unit space instead of applying the Schmidt orthogonal expansion to the unit space as is done when the TS (Taguchi-Schmidt) method is employed. Despite the use of a single-data unit space, highly accurate results were obtained. Because it has fewer unit space constraints, this type of analysis can be carried out on a larger number of signals, and it reduces the economic problems associated with collection of unit space data. Elimination of the major data constraints also clarified the amount of effect on price, which is an essential factor in practical appraisal. The many advantages of this method place it on a par with the TS method.