Mathematical Expression Retrieval Model Based On Ontology And SFE

In the existing mathematical expression retrieval systems,the mathematical expressions being retrieved are not closely related to the semantic documents,in addition,the users have to understand the special format of the mathematical expressions,so they are limited to professionals.For dealing with these problems,a retrieval matching method based on Ontology and SFE is proposed in this thesis after fully studied sequential feature extraction(SFE)and the related technology of ontology.On the basis of using the serialization feature extraction method to analyze the mathematical expressions of the LaTe X format,this thesis constructs the blueprint for semantic ontology library using the authoritative book named mathematics rhe toric,the concept of ontology has been deeply dissected,and the construction scheme of ontology library,which is reasonable and suitable for the semantic retrieval of mathematical expressions,has been designed,furthermore,the relation between mathema tical expressions and their concepts has established using the conceptual modeling tool named Protégé,meanwhile,the semantic ontology library of mathematical expressions is also established.In this thesis,a mathematical expression semantic retrieval model(OSS)has been designed based on Ontology and SFE.According to the advantages of Ontology in querying expansion,the words and expressions of queried words are extended to different degrees,so as to achieve the purpose of retrieving the semantic related literature of mathematical expressions through the input of keywords,concepts,phrases or mathematical nouns.Through the experiment analysis,combined with ontology technology,the OSS realized the semantic extension of the query result set,and according to the mathematical expressions of the LaTe X format and keywords users input,it can accomplish the retrieval of mathematical expressions and related literature using the mathematical expressions which have already built the concept relationship in o ntology library.Compared with the SFE retrieval system,this method improves the value of precision,recall and F1-measure in varying degrees,and it keeps a high rate of expansion when there is a certain retention rate.