Research On Automatic Ontology Construction Based On Relational Database

As a modeling tool which could describe concepts on both of semantic and knowledge layers, ontology has been widely used in knowledge engineering field. It solved the two problem in the developing of knowledge engineering: knowledge reusing and knowledge sharing. Unfortunately, building a high-quality ontology by hands is a complicated and mistakable work, so that the period of building a ontology is long and the cost is high. The paper is to find a method for automatic ontology building, which can use the database resources, to improve the effigency of ontology building.After researching ontology and it's related concepts and analysising existed ontology build method, the paper bring out a new ontology build method which improved the datatable-to-one-concept way in existed method to the datatable-to-semi-finished-ontology way .Based on the new method, the process of ontology building is divided into two main parts: semi-finished ontology extraction and semi-finished ontology merging. Then the paper researched the two parts detailed.In the semi-finished ontology extraction, Formal Concept Analysis is used to analysis the tuples of datatable so that the hidden concepts in the datatable and their relation can be found. In the semi-finished ontology merging,a hierarchical cluster method based on concept similarity is used for merging top-level concepts of semi-finished ontology to generate a hierarchical concept structure,then the lower-level concepts is adjusted for eliminating semantic conflicts to finish the semi-finished ontology merging.Finally, based on the paper had already done, a Ontology Generater Based on Database(OGBD) was established. Then the function modules and their running process were introduced detailed.