Research On Evolutionary Relation Extraction For Domain Knowledge In Web

Because of the rapid development of Internet, it has brought a lot of unstructured and isomerization data, there are many relations among different concepts in the same field of knowledge from Internet, but the knowledge concept of data was disordered and mixed together, for learners it’s too difficult to quick and comprehensive understanding and learning domain knowledge, at the same time, it’s also difficult to construct knowledge in the field of Internet application challenges. Current research on domain knowledge is mostly focused on the up and down a relationship, the r elationship between equivalent relation research. But the evolutionary relation is very important for learners to learn and understand the domain knowledge as well as sorting out the successive logic relationship between two different concepts. But in terms of the present study, has not found the study of this aspect. Therefore, in view of the evolutionary relation between domain knowledge extraction has important research significance and practical value. In this paper, the domain knowledge evolutionary relation extraction under the Web environment has carried out the following work:First of all, in consideration of the particularity of Web data, the method of using word frequency statistics and artificial selection to build key dictionary, which provides the premise to the text classification for Web data;Second, on the basis of the key dictionary building, put forward a kind of Web domain knowledge text categorization method. The experimental results show that the method has good accuracy and recall rate, our proposal is also effective from the Web containing knowledge in the field of data classification in the data, which is the foundation research of the domain knowledge evolutionary relation extraction under the Web;Third, gave in this paper, we study the evolutionary relation between the definition, according to the definition of the evolutionary relation, established the relationship between the evolution of domain knowledge reasoning model, for different structure, different semantic relations of do main knowledge expressing establish accurate syntactic analysis mechanism, using the semantic role of relationship between knowledge concept design domain knowledge evolutionary relation model;Fourth, in the evolutionary relation and evolutionary reasoning model on the basis of the research, put forward an evolutionary relation extraction method on the basis of the Conditional Random Fields,(CRF) model, for the evolutionary relation between different models, construct a unified evolutionary relation extraction theory model. Experiments shows that our method has higher experimental considerations performance than the existing models and our proposal is effective against evolutionary relation extraction for domain knowledge;Finally, on the basis of previous research, the application of the results of evolutionary relation extraction in the field of "machine learning", was designed and implemented a "machine learning" domain knowledge map, the application which can represent the evolution structure of the know ledge-object collection and identify important and difficult points of knowledge effectively, and also can clear display of "machine learning" in the field of domain knowledge with evolutionary relation, to the learners’ learning and understanding of the f ield of "machine learning". this method also has a certain reference value to the subject construction and the related course teaching.