著者

蛭田 有希 浅見 泰司

出版者

公益社団法人 都市住宅学会

雑誌

都市住宅学 (ISSN:13418157)

巻号頁・発行日

vol.2018, no.102, pp.113-122, 2018 (Released:2019-08-01)

参考文献数

11

---

Hedonic pricing method is one of the most efficient approaches for estimating the price of land and buildings. However, not few analysts are concerned about the existence of multicollinearity problems that may cause the inflation of the variance of the estimated coefficients. Because such analysis deals with objects that exist in a space, it is often the case that two or more variables have strong correlation. Although there are some diagnostics for examining the multicollinearity problem, most of the procedures, which are based on the information obtained from the explanatory variables, can only detect the severity of the problem. In terms of detecting the risk of the multicollinearity problem, it is critical that we develop diagnostics incorporating the probability of the occurrence of the problem as well as its severity. This paper suggests a numerical approach for diagnosing the effect of the multicollinearity problem in terms of both probability and severity.