A graph-based approach for representing addresses in geocoding. (March 2023)

• Record Type: Journal Article
• Authors: Zhang, Chen; He, Biao; Guo, Renzhong; Ma, Ding
• Is Part Of: Computers, environment and urban systems. Volume 100(2023)
• Publication Date: 2023-03
• Subjects: Address structure -- Address database -- Address matching -- Toponym disambiguation -- Graph model -- Geocoding; City planning -- Data processing -- Periodicals; Regional planning -- Data processing -- Periodicals; 303.4834
• DOI: 10.1016/j.compenvurbsys.2022.101937
• Languages: English
• ISSNs: 0198-9715

Abstract

Abstract: Addresses, one of the most important geographical reference systems in natural languages, are usually used to search spatial objects in daily life. Geocoding concatenates text with georeferenced coordinates and is an essential middleware service in geographic information applications. Despite its importance, geocoding remains challenging with only text as input, hindering text matching in reference databases without the specific text. To optimize the storage and retrieval of addresses in databases, this work proposes a graph-based approach for representing addresses. The approach clarifies the characteristics of relative concepts, designs a graph structure and identifies modelling strategies. Furthermore, a schema is proposed to perform address matching and toponym disambiguation using an address graph. The model is implemented on a graph database, and experimental tasks are employed to demonstrate its effectiveness. The approach provides a new reference for developers when creating address databases. Graphical abstract: Unlabelled Image Highlights: A graph-based approach for representing addresses is proposed. Manage the spatial relations in addresses separately as a kind of node. Present an address matching method using a joint node and path query on an address graph. Present a computational method of toponym disambiguation based on the in-degree and path length. The approach provides a new reference for developers when creating address databases.