Requirements Auto-generat Model For Domain Ontology Evolution Based On Text

As the key technology of the Semantic Web, ontology need to make timely adjustments according to external knowledge, that is ontology evolution, then it can adapt to dynamic changes of outside. Ontology evolution is to modify all relevant parts of the ontology according to the requirement, how to automatically generate evolution requirement become one of the hot research of ontology evolution.This paper presents a text-based automatic generation model of ontology evolution of the demand, the model can realize the field from the unstructured text to extract relevant concepts in the field, through concept reduction, concept simplification and concept disambiguation operation, then get the key concepts of domain ontology, and then use ontology search algorithms to get the relationship that between selected key concepts and requested evolution domain ontology, finally, automatically generate rules and demand for ontology evolution, according to a pre-defined demand.The model consists of three parts:The first part is text preprocessing. The source of the evolution need to be preprocessed in this part, including speech tagging, non-words and the concept of non-nominal filter; The second part is the key concept screening. Including using of weighted word frequency statistic formula to filter relevant concepts, concept reduction, concept simplification and the concept of disambiguation process. Concept reduction using Porter Stemming algorithm, and semantic disambiguation algorithm using SSI; The third part is automatically generated the demand of ontology evolution. Including searching for the relationship between key concepts and domain ontology, evolution demand generating. Ontology relationship search algorithm is used in this part. Customize the ontology evolution demand generation rule which corresponding to the relationship.This paper designed and implemented ontology evolution demand automatic generation prototype system based on the model, analyzed the experimental data by recall and precision two indicators, verified the prototype system's effectiveness of the key concepts screening and evolution demand generation.