Acta Geodaetica et Cartographica Sinica

Previous Articles     Next Articles

Matching Algorithm for Chinese Place Names by Similarity with Semantics of General Names for Places into Consideration

  

  • Received:2014-01-21 Revised:2013-06-12 Online:2014-04-20 Published:2014-02-18

Abstract:

Matching of Place Names is one of the key issues in geographic information retrieval, integration and update for multi-source geospatial data. According to the constitute characteristics for Chinese Place Names and relations between general names for places and its types; Ontology knowledge base for general names for places has been established, based on which semantic of Place Names is used as an important indicator for matching of Place Names by their similarity. Aiming at overcoming the shorts of queries by literals or geospatial data for place names, we propose a new matching algorithm and query strategy for Chinese Place Names taking both similarities of special names and general names for places into consideration. The method simulates human cognitive habits, in which the weights of similarity for special names and general names for places are set dynamically according to the degree of semantic similarity by a monotonic function relationship. The final composite similarity index for Place Names is weighted average for similarities of special names and general names for Places. Based on the model above, the strategy and flow has been put forward, which enhances the theoretical basis and completeness of matching algorithm for Chinese Place Names by using semantic knowledge from general names for places, and it thereby improves the accuracy of the matching algorithm. The experimental results show that the matching model is consistent with human cognitive habits, and further demonstrate the rationality and effectiveness of this method.

Key words: general names for places, semantic, ontology, composite index, matching of place names