Research On Knowledge Discovery Based On Fuzzy Ontology Fusion And Reasoning

Knowledge discovery is an eternal topic in the progress of human civilization.With the development of the era of big data,there are various kinds of knowledge resources in various platforms of the Internet.Data mining,knowledge fusion,rule reasoning,neural network and other related technical means are adopted for these multi-source and heterogeneous knowledge resources,which has a great role in promoting knowledge discovery.With the development of Web 2.0 to the era of web 3.0,all kinds of knowledge resources on the Internet will be transformed to the form of relational data and RDF triple.Knowledge discovery activities will rely more and more on its "relational" knowledge resource environment,so knowledge discovery will inevitably face new opportunities and challenges.At the same time,due to the complexity of network knowledge resources,knowledge is often both accurate and uncertain.Ontology is a formalized,clear and detailed description of the conceptual relationship system that can effectively express knowledge and provide support for users in the process of knowledge discovery.Therefore,from the perspective of knowledge ambiguity,this study considers the coexistence of knowledge accuracy and ambiguity,constructs a fuzzy ontology containing precise knowledge and fuzzy knowledge,and explores the technology and implementation process of knowledge discovery in ontology environment,which will help to promote the development of knowledge discovery in the era of Web 3.0 and provide ideas for related research.Firstly,this thesis introduces the research status of knowledge discovery at home and abroad,discusses the content of this study,and confirms the research methods and innovations of this thesis.Then the related concepts and theoretical basis are studied and the current concepts,representation methods and related operations of fuzzy knowledge are introduced in detail.The related theories of Ontology-based Knowledge Fusion and reasoning are discussed,and the object and process of knowledge discovery are described.Then,based on the theory of fuzzy knowledge and OWL ontology description language,a new model of fuzzy ontology representation is proposed.The related technologies of building,fusing and reasoning of fuzzy ontology based on this model are studied.Finally,we proposed a knowledge discovery model based on fusing and reasoning of fuzzy ontology.It is found that the OWL-based fuzzy ontology model proposed in this thesis can describe both precise knowledge and fuzzy knowledge..The proposed fuzzy ontology is constructed by crawling relevant data of pharmacokinetics of drug-drug interactions from different network knowledge resources.The fuzzy ontology from different knowledge resources is fused to realize the construction of global fuzzy ontology.Then,reasoning rules are constructed according to drug similarity mechanism and pharmacokinetic action mechanism to complete drug interaction knowledge discovery.It is found that the fuzzy ontology model proposed in this thesis based on OWL ontology description language can describe both precise knowledge and fuzzy knowledge.It can directly complete ontology-based precise reasoning and fuzzy reasoning.It does not need to extend and transform the language,and can also complete both precise knowledge discovery and fuzzy knowledge discovery.This experiment combines the fuzzy rules of drug similarity and the precise rules of pharmacokinetic mechanism to discover the knowledge of drug-drug interactions.Compared with the previous research on drug-drug interactions knowledge discovery based on rule reasoning and pharmacokinetic mechanism,the accuracy of the experimental results in this thesis has been reduced,but the recall rate has been greatly improved.However,knowledge discovery of drug interactions based on pharmacokinetics has limitations,because the actual drug interactions and their potential mechanisms involve complex pharmacological processes.In addition,because the database has not kept up-to-date updates and more drug interactions have not been found,it is impossible to prove whether the interaction drug pairs in the test samples that are not known by reasoning really do not have drug interactions,so the recall rate experimental indicators are more important in this experiment.So researchers can conduct targeted clinical experiments to find out whether the two drugs interact with each other,which helps to save resources and avoid blind discovery.At the same time,the correctness and validity of the knowledge discovery model based on fuzzy ontology fusion and reasoning in this thesis are verified,which can be used for reference in the research of knowledge discovery in the era of web 3.0.