Description of Application based on the Alexandria Digital Gazetteer Protocol
Martin Doerr
and Manos Papagelis
Institute of Computer Science, Foundation for Research and Technology
- Hellas
P.O. Box 1385, GR-71110, Heraklion,
martin@ics.forth.gr
papaggel@ics.forth.gr
Department of Computer Science, University of Crete
P.O. Box 2208, GR-71409, Heraklion,
Introduction
Most of the information available digitally, as well as, information in library catalogues, bibliographic indexes, and museums has some reference to locations or objects in geographic space. However, only a fraction of it is properly georeferenced, i.e. indexed using a geographic coordinate system. In most cases, simple place names and other non-formalized descriptions are used. Besides, between data sources different place names (variant place names) are used to describe identical places. The key to merging these data sources with the geospatial data and thus toward the promise of the Digital Earth metaphor for information management is the digital georeferenced gazetteers.
Digital georeferenced gazetteers link place names to geographic footprints. Current digital library implementations (for example, the Alexandria Digital Library and the Digital Library Project of the University of Berkeley) have demonstrated the use of such gazetteers to provide indirect georeferencing to geospatial datasets. Through gazetteer translations, a place name can be converted to a geographic footprint that can be used to find relevant geospatial datasets (Indirect georeferencing).
However, there are instances of geographic names that are non-unique within geographic domains. For example, the geographic name “Athens” is not identified by its name in the global geographic domain, so if we query a digital georeferenced gazetteer using the geographic name “Athens” the response will consist of more than one results, however only one of them is correct according to what we had in mind. It is obvious that when integrating georeferenced information we need a one-to-one correct mapping between a vague geographic name of a datasource and a recognized and identified geographic name in a digital gazetteer, otherwise the integrated information will suffer from essential semantic problems and the benefits of the unified information will be reduced.
Application Description
Data cleaning, precision and recall, identification and verification are some of the aspects that are encountered throughout this work and the challenge is the automation of the mapping process surpassing the misspelling, incompleteness and duplicate detection problems that are involved, so that human intervention can be systematically reduced.
Our main objective is to analyse aspects of correctness and completeness of digital gazetteers and to stress the impact that they have to the mapping process. Based on our study, we propose methods that automate the mapping process. We aim at the maximum accuracy of the mappings so as to eliminate the human control afterwards. Ideally, we would like our algorithms to be able to decide with one hundred percent probability (100%) that an identification is correct. The goal is that our algorithms provide considerably smaller amount of the total data to the user that need to be manually controlled. We believe that our work demonstrates a relevant way in which digital gazetteers could be used in order to add value in information integration projects and especially where geographic names serve as critical point of properties linkage between heterogeneous resources.
In our application, we try to identify the place names that exist in diverse datasources by using a third-party digital georeferenced gazetteer, the Alexandria Digital Gazetteer. For each single place name that is described in a datasource, we try to form a query that follows a standard format. This query includes information about the place name, wider place names, if available, that include this place name and finally, the country in which this place name falls within. Each query evaluation results in an answer string that follows a standard format. The application takes as input a set of query strings and forms the answer strings by querying the Alexandria Digital Gazetteer using the Gazetteer Protocol. We use a portion of the ADL query language types such as identifier-query, name-query, class-query, relationship-query, footprint-query with a combination of available operators in order to generalize or to specialize our query. Below, there are more formal descriptions, as well as, examples of query and answer strings.
Query string format
Format
Place/WiderArea/…/WiderArea@Country
where,
Place = PlaceName(PlaceAlternative1;…;PlaceAlternativeN)
WiderArea = WiderAreaName(WiderAreaAlternative1;…;WiderAreaAlternativeN)
Country = CountryName
Explanation
PlaceName: This is the name of the place
PlaceAlternative(1..N): Alternative names of the PlaceName
WiderAreaName: Wider area on which the PlaceName falls within.
CountryAlternativeName(1…N): Alternative name of the WiderAreaName
Description
To form a query string we first write a PlaceName. If we have knowledge of alternative names of this PlaceName then we include in parenthesis [()] the PlaceAlternative names. PlaceAlternative names are separated by semi colon [;]. As you can notice the WiderArea is Recursive. We can have as many as possible WiderAreas of a place. We call them levels. Levels are separated by the “slash” symbol [/]. When we have described the Place and the WiderAreas of it we write the “at” symbol [@] and then write the CountryName.
Examples: