Educing spatial knowledge from educational resources: semantic information extraction of concepts and places

Journal of Spatial Information Science

Margarita Kokla ; Eleni Tomai ; Vagelis Papadias

关键词: semantic enrichment;    semantic information extraction;    ontology;    ontology-based information extraction;    unstructured texts;    spatial concepts;    place names;    named entity recognition;    visualisation;   

学科分类：计算机科学（综合）

来源: University of Maine

【 摘 要 】

Semantic enrichment aims at enhancing content interlinkage, search, and discovery by adding well-defined semantic metadata that help machines make sense of the content. Information extraction plays a central role in this process since it supports the automatic identification of specific types of information from unstructured or semi-structured natural language text. An emerging subfield of information extraction is ontology-based information extraction, in which the process is guided by an ontology as a means to formally describe domain knowledge and assist the extraction of pre-defined concepts, properties, and instances. The proposed approach implements a semantic information extraction process in order to exploit a rich source of spatial knowledge: educational resources which provide a wealth of spatial knowledge in terms of abstract spatial concepts and place names. Place names are extracted using named entity recognition combined with spatial / administrative hierarchy of place names from GeoNames to extract alternative locations and support the enrichment and geovisualization of educational resources. Concepts are identified using extraction rules in the form of regular expressions and lexico-syntactic patterns. An ontology with a natural language anchorage is used to guide the extraction and semantic visualization process supporting inference rules to further extract explicit knowledge in a synergistic approach where extracted information is used to enrich and enhance the ontology. Evaluation by means of manual semantic information extraction provides a first validation of the approach thus far.

【 授权许可】

CC BY