Font Size: a A A

Research On Multi-source-oriented Domain Ontology Construction Based On Formal Concept Analysis

Posted on:2011-06-12Degree:MasterType:Thesis
Country:ChinaCandidate:Y J PanFull Text:PDF
GTID:2178330338476273Subject:Computer application technology
Abstract/Summary:PDF Full Text Request
In order to shield the difference between Knowledge Systems, Ontology's introduction is an irresistable trend. It began a new train of thought to solve the knowledge sharing and knowledge reuse. And also provided support for machine to understand semantics automatically. This paper did research on the construction method of domain Ontology from home and abroad, then proposed a multi-source-oriented Ontology construction method based on FCA according to the shortcomings such as data source unicity, low-level automation and construction method complexity of foregone methods, and analyzed some important relative algorithms about this method.In this paper, the Ontology is viewed as a style of trinity-structure: O = {C,R,A}, which include C(concepts), R(relations) and A(axioms), and carried out the trinity-structure's extraction with the result of finishing the Ontology construction.(1)As for the concept extraction, this paper introduced the related theory of FCA so as to low complexity and increase construction efficiency. Firstly used the different methods to get its formal context according the different data source, then merged and splinted these contexts to the independence contexts. Finally created formal concepts and concept lattice with the application of an algorithm of generating concept lattice based on rank of matrix. This lattice is visualized by Hasse graph.(2)As for the relations acquisition, this paper used the formed concept lattice to extract the semantic relations and promoted an improved concept similarity model to obtain the non-semantic relations in order to improve the precision. The experiment data proved the model has a good practicability.(3)For the axiom extraction, this paper introduced the Association Rules Mining algorithm of Data Mining to perform this task.This paper used the digital products as example, which applied the Ontology construction method, and realized some relative algorithms of the method involved, finally gave the Conclusion of the method in the view of the whole situation. At last conclude the work of the paper and prospect the future and the train of thought studied further.
Keywords/Search Tags:Ontology, FCA, concept, concept lattice, relation, axiom
PDF Full Text Request
Related items