A KEYWORD RETRIEVAL AND SORTING METHOD BASED ON ONTOLOGY FOR ENGINEERING DATA

Zheng Zhang,Weidong Yang,Xinyue Li,Bole Shi

DOI: https://doi.org/10.3969/j.issn.1000-386x.2014.11.011

2015-01-01

Abstract:Engineering data exist widely in modern manufacturing industries.The keyword query method traditionally adopted in engineering data is hard to deal with heterogeneous documents and to identify the semantics of the keywords in engineering sector.Hence, we propose a new keyword retrieval framework for engineering data to tackle this deficiency.This framework extracts the text information from various heterogeneous engineering documents with proper tool first by building an ontology which represents the engineering data, and then sets up the“XML snapshots”.The XML snapshot tags the corresponding otology concepts to the indexed keywords.When a query comes, the framework first matches the keywords queried by the user to ontology concepts, and we present an algorithm to abate the semantic confusion of the keywords in matching process.On this basis, we then present a sorting model, which marks and sorts the XML snapshots tallied with the query, the marking emphatically considers the extent of the snapshot satisfying the keyword query in terms of semantics.The retrieval framework finally sorts and returns the original documents corresponded to the snapshots to user, makes the heterogeneous engineering data documents be able to processed and queried in uniform.This framework has high recall and precision ratios on retrieval results of engineering data, and has reasonable retrieval results sorting as well, it is suitable for massive engineering data documents in practical enterprises, and overcomes the defects of traditional engineering data retrieval.In experiment the actual datasets are used, it verifies and illustrates its effectiveness in practice and efficiency in performance.