An Intelligent Retrieval Algorithm For Enzyme Ontology Based Semantic Extension

With the development and the application of the information technology,the keyword-based information retrieval technology can’t meet the users’ needs of the knowledge and semantic.Thus finding the new ways has become a research focus.The ontology learning and ontology technology can tap the implicit knowledge from the corpora text making its processing and reuse.Facing the different structures and large amounts of text data, extracting knowledge is difficult but very significant.Which information retrieval integrates with semantic knowledge can make the computer having a certain semantic understanding,improving search efficiency and user satisfaction.This research puts forward the semantic retrieval framework based on ontology learning,and discusses the related intelligent algorithm.Surrounding the semantic retrieval framework, the thesis uses the statistical method to do the ontology learning.On the basis of reviewing the related literature and participating with the domain expert,this paper sums up that the expression of different domain knowledge in its literature is not in the same law.According to the domain knowledge characteristics of the enzyme,this research finds the logical and efficient way to extract knowledge and explores how to make use of knowledge in information retrieval.This method can convey the retrieval intention fully,thus the retrieval efficiency and satisfaction are increased.First,according to the characteristic of enzyme domain knowledge, this approach proposes the extraction method of concept and relationship, and filters the results of the extraction by semantic filtration. The method of the ontology learning can improve the accuracy of ontology learning and the efficiency of constructing domain ontology. Second,the artificial bee colony algorithm search the text space forthe concepts and relationships, this can have a strong global convergence, stepping out the local optimum and increase the diversity of solution. Finally, this paper proposes the strategy of the semantic expansion. On this basis, the semantic full-text retrieval is achieved. The experimental results show that this framework can improve the efficiency of semantic process in information retrieval.The innovation of this study lies in the following: this paper proposes the ontology learning method based on statistics. The relevance of the domain is defined by comparing the domain corpus and compositive corpus. Then according to the characteristics of the enzyme domain knowledge, the knowledge about a particular enzyme is emerged in a range of related field in corpus and the other terms which appear uniformly are the predicate of the domain terminology. The domain determinacy is defined. The domain relevance and the domain determinacy decide the condition of the term extraction. The semantic filter deal with the results of the extraction, this can improve the correctness of the ontology learning and the productiveness of the ontology construction.The artificial bee colony algorithm retrieves the domain text corpus for term and relationship. In order to avoid that the follow bee selecting the optimum nectar greedily, it adds the search directionality to the follow bee and lead bee’s retrieval.The design of the sensitiveness can expand the search area with the fitness of the optimizational problems.Then ABC algorithm can avoid falling into local optimum.The parameter of scouts bee’s behaviour is difficult to control, so this paper makes the scouts bee search the area nearby the keyword in the text space firstly. This can prevent the blindness of the search.On the basis of the previous work, the paper discusses and realizes the semantic framework of full-text retrieval. This framework associate the semantic with full-text retrieval, including the ontology learning, ontology construction, semantic full-text retrieval and the evaluation of the results. The results can affect the parameter of ontology learning algorithm.The significance of this study lies in the intelligent semantic retrievalframework,which makes the information retrieval system has a certain ability to learn and understand the semantics.After the ontology learning,the concept terms and relationships related to domain knowledge are used to build ontology with hierarchical structure of the specification.And also the full-text retrieval uses the knowledge ontology to understand the semantics of the query words.Finally,the evaluation of retrieval effectiveness can adjust the parameters of the ontology learning algorithm,which improving the learning and retrieval capabilities.