Research On Representation And Application Of Tea Knowledge Ontology-based

In the Internet era,knowledge is presenting a massive,multi-source,isomerization trend with rapid development of information technology.How to organize and manage knowledge effectively is being a hot topic in the field of information retrieval.Ontology,as a new tool of knowledge organization,has a good representation of semantic relations and the characteristics of supporting logical reasoning,and is widely used.Tea is one of three nonalcoholic beverages in the world and has growing areas around the world.China,as the birthplace of tea,has a long history of studying tea.Tea knowledge involves numerous subject fields,such as Cultivation,Biochemistry,Pests and Diseases,Inspection,Mechanology,Cultural Customs,Industrial Economy and so on.In this context of technology and knowledge,this paper regard the rich tea knowledge as the study object and realize knowledge organization and retrieval system by ontology technology.This paper is mainly divided into three parts:In the first part:firstly,this paper studies definition,classification and application of ontology.Then the paper gives in-depth understanding of development on organizational tool in knowledge economic society and points out that ontology is especially focused on due to comparative analysis the advantages and disadvantages of other information organizational tools.At the same time,this paper makes investigation and analysis of research status on agricultural ontology.Finally,this paper accumulates fundamental knowledge of ontology construction such as construction method,editing tools and development tools so as to do learning for subsequent tea domain ontology construction.In the second part:the author finds ontology construction artificially time-consuming,strong relying on experts so decides building tea domain ontology semi-automatically.Then,with depth analysis of ontology learning methods,this paper uses the comprehensive method of "seven step" and "skeleton method" to construct tea ontology.First the method uses the ICTCLAS segmentation system to segment word and tag POS according to obtained data,writes a program to complete specified POS and stop words deleted.Then using the TF-IDF method to extract tea concepts based on feature weight in order to candidate concept set.And the set is standardized and complemented combined with thesauri,dictionaries and experts.Third,this paper sets the values of support and confidence threshold with association rule mining method to identify the relationship between concepts.Through the above steps,We can obtain the corresponding class,attribute,example of tea ontology,using Protege to complete the ontology formal representation including class hierarchy confirmed,object attribute domain and range set,data attribute restricted and so on.Last,the writer adds the step of ontology evaluation and optimization by using HermiT inference engine of Protege to check the consistency of logic thus to prove the rationality of tea domain ontology.In the third part:this part tells the application of knowledge retrieval based on tea ontology.First,it elaborates the limitation of traditional information retrieval containing user faithful expression,word matching,lexical island and the advantage of knowledge retrieval involving semantic matching and intelligent reasoning.Second,the paper discusses the solution of key technologies on tea ontology knowledge retrieval,including the function of query expansion,information indexing,resources retrieval.Specifically,It uses Jena semantic package to read and parse of ontology.Last,the paper develops retrieval system by Eclipse to achieve semantic expansion of synonyms,hypemyms,relatives in keyword retrieval method so as to improve the degree of recall and precision.