2026-07-25T03:01:02Z https://drops.dagstuhl.de/oai

oai:drops-oai.dagstuhl.de:20827 2024-09-09T09:21:19Z ddc:004 open_access

Exploring Discrete Spatial Heterogeneity Across Quantiles: A Combination Approach of Generalized Lasso and Conditional Quantile Regression (Short Paper) Inoue, Ryo Aoki, Kenya discrete spatial heterogeneity generalized lasso conditional quantile regression Spatial heterogeneity has been investigated extensively. However, in addition to spatial heterogeneity, there are spatial phenomena where heterogeneity in the data generation process exists across quantiles. This study proposes a new method that combines generalized lasso (GL) and conditional quantile regression (CQR) to analyze discrete spatial heterogeneity across quantiles. GL enables the identification of spatial boundaries where the spatial data generation process varies discretely, and CQR estimates the parameters of the conditional quantile of the dependent variable. The proposed method is expressed as a linear programming problem and is simple to use. To validate its effectiveness, we applied this method to apartment rent data in Minato Ward, Tokyo. The results revealed that the neighborhoods where rent levels deviated from the overall trend in the analyzed area differed by quantiles. Schloss Dagstuhl – Leibniz-Zentrum für Informatik Ryo Inoue and Kenya Aoki 2024 Is Part Of LIPIcs, Volume 315, 16th International Conference on Spatial Information Theory (COSIT 2024) InProceedings Text doc-type:ResearchArticle publishedVersion application/pdf doi:10.4230/LIPIcs.COSIT.2024.12 urn:nbn:de:0030-drops-208272 https://drops.dagstuhl.de/entities/document/10.4230/LIPIcs.COSIT.2024.12 eng https://creativecommons.org/licenses/by/4.0/legalcode