| Co-Investigator(Kenkyū-buntansha) |
DOHI Tadashi Hiroshima University, Graduate School of Engineering, Professor, 大学院・工学研究科, 教授 (00243600)
SUZUKI Atsuo Hiroshima University, Graduate School of Engineering, Professor, 数理情報学部, 教授 (70162922)
FUSHIMI Masanori Nanzan University, Department of Mathematical Sciences, Professor, 数理情報学部, 教授 (70008639)
OKAMURA Hiroyuki Hiroshima University, Graduate School of Engineering, Associate Professor, 大学院・工学研究科, 助教授 (10311812)
|
|---|
| Research Abstract |
We have developed a mathematical model of evaluating fixed-assets for housing and made a system based on the mathematical model. In particular, the mathematical model has been extended to the multiple regression analysis form the quantification method I which is used in the previous researches. In the quantification method I, we have to categorize quantitative measures for pricing land such as the distance to station to several classes, so that the. quantitative measures would be regarded as discrete values. By using the multiple regression analysis, we have been able to use the quantitative measures directly to evaluate fixed-assets for housing. Similar to the multiple regression. analysis, the evaluation method on the Cox regression analysis has been proposed, where the Cox regression analysis is performed in a proportional hazard model. The proposed method takes an advantage to estimate the range of land prices, which is the interval estimation of land price, because the land price is represented as a random Variable in the proportional hazard model.
In numerical experiments, we have performed the multiple regression analysis and, the Cox regression analysis under the measures pricing land for housing which are collected in Higashi-Hiroshima City, and have compared two analyses quantitatively. To evaluate the goodness-of fit to the real land price for the multiple and Cox regression analyses, we use mean absolute errors, likelihood, AIC, BIC. As a result, the Cox regression analysis is superior to the multiple regression analysis in terms of pricing land for housing. In addition, the evaluation system has been developed on spreadsheet software, where we can estimate the prices of land with simple operations. In this system, the recent estimation methods are applied for making the fast estimation. Finally, we will submit the paper on this research to the journal of industrial and applied mathematics in Japan.
|
