The Relevance of Spatial Relation Terms and Geographical Feature Types Reporter Chunju Zhang Date: 2012-05-29
1 2 Introduction Classification of Spatial Relation Terms 3 4 5 Calculation of Relevance Conclusion
The semantic research of spatial relations is the premise and basis for the description and expression of spatial relations. Spatial relation terms generally indicate spatial relations described in natural language context. describes the nature of people s internal representation of space and is the primary means for representation and exchange of geographical information, such as geographical entities, spatial relations, etc.
mountains waters The semantic representation patial relation termsis closely related to geographical entities and their characteristics e.g. geometry, scale and geographical feature types. watershed some spatial relation terms can be used for several different geographical feature types, while some are just for a certain geographical feature type.
geographical information retrieval GIS natural language query semantic meaning of spatial relation terms qualitative spatial reasoning...
Classifica3on of Spa3al Rela3on Terms Spatial relations in natural language are richer, qualitative, fuzzy, uncertainty and unstructured, while spatial relations in GIS are quantitative, structured, and accurate. In different language, spatial relations terms are with different diversity and complexity. of (hengguan) (hengguan) GIS (chuanyue) cross (hengguan) (jiaocha),
Classifica3on of Spa3al Rela3on Terms sed on RCC8, TPCC the frequency of spa8al rela8on terms in natural language context basic categories of spa8al rela8ons and classifica8ons of spa8al rela8on terms are described
Calcula3on of Relevance This paper takes the large scale annota8on corpus (Geocorpus) of spa8al rela8ons of Chinese Encyclopedia (geography) and 600 commonly used spa8al rela8on terms as an experimental data to explore the seman8c relevance of spa8al rela8on terms and geographical feature types of geographical en88es in text. A binary spa8al rela8on could be formalized as <geographical en3ty A, spa3al rela3on terms, geographical en3ty B > in natural language context. <feature type of geographical en3ty A, spa3al rela3on term, feature type of geographical en3ty B >.
Calcula3on of Relevance Overlap is a classic calcula8on method for seman8c rela8ons, and it is based on the co occurrence frequency of two events in a data set T represents the occurrences of a spa8al rela8on term in the Geocorpus; A ' and B' denote the occurrence of two geographical feature types; R indicates the relevance degree between spa8al rela8on term (T) and a pair of geographical feature types (A 'and B').
Calcula3on of Relevance Taking the spatial relation term of (liuru, flow into) as an example, the results of the relevance are just as shown in following figure. river water system channel Other water system canal sea reservoir A single geographical feature type Water course riverside River head drainage area Marine elements drainage basin 0.667 lake Flow into ocean pond delta gulf.. outfall A pair of geographical feature types the relevance of spatial relation terms and Geographical feature types
S Ontology formally represents rich knowledge as a set of concepts and the rela8onships between those concepts within a domain. A knowledge base is developed based on protégé to formally represent and visualize seman8c knowledge of spa8al rela8on terms. geographical feature types the value of relevance spatial relation terms spatial relation classifications
Conclusion we However, the annotation quality of the corpus and the classification granularity of geographical entities have a great effect on the performance, especially for a general dataset. the relevance of spatial relation terms and the type of geographical entities have big limitations based on manual annotation and judgment completely.
ANY QUESTIONS