search for




 

Prediction of the land price based on deep learning and residual kriging
Journal of the Korean Data & Information Science Society 2021;32:475-85
Published online May 31, 2021;  https://doi.org/10.7465/jkdi.2021.32.3.475
© 2021 Korean Data and Information Science Society.

Jinho Choi1 · Yongku Kim2

1Korea Real Estate Board
2Department of Statistics, Kyungpook National University
Correspondence to: 1 Research fellow, Korea Real Estate Board, Daegu 41068, Korea.
2 Corresponding author: Associate professor, Department of Statistics, Kyungpook National University, Daegu 41566, Korea. E-mail: kim.1252@knu.ac.kr
Received March 29, 2021; Revised April 15, 2021; Accepted April 19, 2021.
This is an Open Access article distributed under the terms of the Creative Commons Attribution Non-Commercial License (http://creativecommons.org/licenses/by-nc/3.0) which permits unrestricted non-commercial use, distribution, and reproduction in any medium, provided the original work is properly cited.
Abstract
This study is propose the conjunction models of deep neural network and residual kriging (DNNRK) for advancement of land price estimation and validate its performance. Seoul (Gangnam 3 District) was chosen as a study area and we collected land prices of standard lots estimated by appraiser. When inspecting the model performance based on the test data, we confirmed that the land price accuracy form the DNNRK model was improved substantially compared with that of the OLS, DNN and RK. Therefore, these results indicate that introducing spatial autocorrelation as a location factor to models can improve the performance of land price prediction significantly. Furthermore, DNNRK model coupling DNN and residual kriging can be an effective alternative for estimating the land price accurately.
Keywords : Deep neural network, land price, residual kriging, spatial autocorrelation.