Font Size: a A A

Research On Construction Of Knowledge Base And Knowledge Discovery Of Traditional Mongolian Medicine Based On Domain Ontology

Posted on:2019-01-21Degree:DoctorType:Dissertation
Country:ChinaCandidate:Y L BaoFull Text:PDF
GTID:1364330548962762Subject:Library and file management
Abstract/Summary:PDF Full Text Request
We have entered the “data-driven” of “wisdom age” from the information age,and the resource utilization of data has become an important direction of digital library.The resource utilization of data will become the development direction from digital library to knowledge service.It is also a trend that the demand for semantics and deep mining of library resources will provide discovery services which focus on problem-finding and users-discover for information.It can extract knowledge from the library resources which is hidden based the specific demands of users,and provide users with the information which are understandable and usable to help users to analyze and deal with problems.With the concept of Semantic Web proposed,ontologies with semantic description capabilities have gained extensive attention.Ontology technology is a common knowledge representation method of data-semantics.Its basic idea is to represent domain knowledge as a labeled graph,where nodes represent domain concepts and edges represent the semantic relationships between concepts.Semantic networks have been widely used in many fields such as computational linguistics,biology,and medicine and so on,its greatly due to its advantages of simplicity,flexibility,richness,and readability.The design concept of the Semantic Web has been reflected in Word Net,UMLS,SNOMED CT and other large terminology systems.The ontology technology is used to describe and reveal the semantic relationship between the basic theories of Mongolian medicine,diseases,symptoms,prescriptions,medicinal materials,medicinal properties,medicinal tastes,and diagnosis and treatment methods.Building a knowledge base,which is an effective way to achieve the semantic retrieval,semantic inference,and knowledge discovery of Mongolian medicine resources.This article selected important Mongolian medical literature,including authoritative reference books,ancient Mongolian and Chinese books,modern books,journal articles,dissertations,and other data sources,then established a basic digital text set of Mongolian medical science.According to the characteristics of Mongolian medical science,with reference to the International Standards Semantic Network Framework of Traditional Chinese Medicine Language System(ISO/TS17938:TCMLS-SN),a Mongolian-Chinese bilingual Mongolian medicine concept semantic classification hierarchy model and semantic relationship model are established.According to the semantic model,three-layer neural network model,word embedding and KNN classification algorithm are applied to classify and tag the concept of basic digital text set of Mongolian medicine.Then,a basic concept base of Mongolian medicine is established,and the domain ontology of Mongolian medicine will be constructed,and developed Mongolian medicine knowledge base prototype system.The main research includes:(1)Preprocessing of Mongolian medical literatureAccording to the recommendations of experts in the field,they selected important ancient books,modern writings,authoritative reference books,journal articles,and dissertations as the data sources.The digital texts of the data sources were collected using full-text databases such as the Mongolian Ancient Books Database,the Mongolian Modern Books Database and the Chinese Basic Ancient Books Database established by the Inner Mongolia University Library.Print OCR identification and proofreading of printed documents that cannot obtain digital texts and establish basic digital text sets.(2)Mongolian medical domain conceptual system modelMongolian Medicine has its own unique theoretical system.Mongolian medical science is guided by the theory of yin and yang,five elements and five yuan theory,and it permeates the whole view of man and nature.Mongolian Medicine condensed the "five elements"(or five yuan)into the "three-factor theory," namely,Hei,Sheila,and Badagen.The "three-cause theory" is the theoretical basis of Mongolian medicine.It is used to explain all life activities and pathological processes and to guide the practice of diagnosis and treatment.Based on the theory and practice characteristics of Mongolian medicine,referring to TCMLS-SN,the concept of Mongolian medicine is classified from the semantic level,and the concept semantic type and semantic relationship in Mongolian medicine are defined.Define the semantic types of Mongolian medicine,its sources include:(1)Characteristic concepts in the field of Mongolian medicine,such as "three roots","seven elements","six basic diseases","black veins","white veins" and "blood therapy";(2)Conceptual equivalent concepts in the field of traditional Chinese medicine,such as "organs","acupuncture points",etc.;(3)General concepts such as "symptoms","symptoms","etiology","pathogenesis","medicinal substances",etc.At the top level,it is divided into two categories: "Entity" and "Events",and from this,its hierarchy is developed to form the concept semantic model of Mongolian medicine.(3)Mongolian medical document miningApply the NLPIR Chinese word segmentation system of the Institute of Information Technology of the Chinese Academy of Sciences and the Mongolian word segmentation system of the Pattern Recognition and Artificial Intelligence Laboratory of Inner Mongolia University to segment word processing for digital texts,segment basic vocabularies,and establish a basic vocabulary based on the basic thesaurus.This paper proposes a method based on the word vector package for vocabulary classification and semantic annotation to generate a basic concept set of Mongolian medicine.(1)The generation of word vector.According to the semantic types and semantic sets of the Mongolian medical sciences defined in the previous section,the vocabulary recognition in the basic lexicon obtained by word segmentation is classified as one or more of the above semantic categories and semantic relationship sets,ie,all words in the text are marked with a One or more semantic types or semantic relationship tags.Word vector technology is used to express the concept of noun semantics in texts,and a text annotation(classification)recognition model is generated through machine learning algorithm training.(2)Mongolian medical concept semantic annotation model.After the word vectors are generated,the classical classification model k-nearest neighbors(KNN)of machine learning is used to implement the concept of classification tasks.That is,each semanticnoun is classified into one or more types of labels for the semantic types and semantic relations of Mongolian medicine.(3)Set concept of the field of Mongolian medicineThe basic thesaurus has been classified and semantically labeled to form Mongolian-Chinese Mongolian medicine basic semantic concept set.The topic will be optimized through tools such as domain expert consultation and Mongolian Semantic Information Dictionary to form the semantic concept lexicon of Mongolian Medicine and Mongolian and Chinese.(4)Ontology construction of Mongolian medicine based on concept latticeAfter acquiring the basic semantic concept thesaurus of Mongolian medicine,the idea of constructing ontology of Mongolian medicine is: based on the semantic concept thesaurus in the process of skeleton method,the purpose and scope of ontology construction are clear;A top-down approach was used to analyze the domain ontology on the characteristics of Mongolian medicine,and then refer to the Mongolian medical theory system and the assistant of domain experts,the relationship between concepts and the addition of instances were established;finally,the appropriate formal language was used to represent the ontology.(5)Conceptual Semantic Retrieval and Reasoning in Mongolian MedicineOntology construction establishes the foundation for semantic reasoning.The subject will use the SWRL(Semantic Web Rule Language)rule language and Jess inference engine to realize diagnosis reasoning and prescription recommendation based on the fact that the Mongolian medicine ontology provides inference facts.For example,in the practice of Mongolian medical treatment,doctors learn about the condition of the patient by observing the patient,dictating the condition of the patient and in combination with some of the examination results of the current medical science.The conditions of the patient are summarized as symptoms,pulse and tongue and other images of Mongolian medicine.Through these concepts,the patient's syndromes are determined and prescriptions are issued for the syndromes.In the ontology of Mongolian medicine,there are three important conceptual categories and their subordinate concepts: symptoms(including major symptoms,minor symptoms,pulse,and tongue phase);syndromes and prescriptions.(6)Construction of the Mongolian Medicine Knowledge BaseThe development of a knowledge base in the field of Mongolian medicine will be carried out in response to the actual needs of clinical medicine,education,teaching and scientific research in Mongolian medicine.Based on the above researches,combing the system development and the integration of various types of algorithms,this paper will realize the evolution and iteration of ontology based on domain expert intervention.And these researches and developments on the basis of the Jena ontology of HP Lab will make it suitable for storage,query and reasoning of the Ontology field of Mongolian medicine.The knowledge base will cover the semantic types of all Mongolian medicine fields,with the functions of semantic search and reasoning,knowledge visualization,diagnosis aids,misdiagnosis prompts,and so on to realize the transformation from domain literature to domain knowledge.The main innovations of this thesis is as follows:(1)Definition of Semantic Concept Sets in Mongolian Pharmacology Based on Literature Data Mining.Based on the characteristics of Mongolian medicine theory system,this paper defines semantic types and semantic relation sets of the concept of Mongolian medical science at the semantic stratification.And through the literature data mining,the domain concept is classified and labeled,and then the semantic concept set of the Mongolian medicine field is constructed.Domain semantic concept set is not only the foundation of domain ontology construction,but also an important evidence for the standardization of terminology in Mongolian medicine.Proposing a Technical Route for Constructing Domain Ontology of Mongolian Medicine which Based on the Complementary Integration of Concept Lattice and Ontology.Concept lattice technology was introduced in ontology construction of Mongolian medicine.Using the concept of formal and background analysis to construct the domain ontology,it can not only fully reveal the conceptual semantic relationship in the Mongolian medicine field,but also to some extent eliminate the concept ambiguity arising from the non-standard and inconsistent concepts in different literature classics.(2)Mongolian-Chinese Bilingual Ontology Construction Based on Equivalent SemanticsBased on the semantic definition rules of OWL,this paper presents a bilingual construction model of equivalent semantics.The model defines four equivalent semantic forms including the equivalent class,the equivalent object attribute,the equivalent data attribute,and the equivalent individual.Through the hierarchical structure of Mongolian medicine domain concept obtained in the third chapter of this paper,the corresponding Class,sub Class and Individuals are established.Implement the semantic interconnection of Mongolian and Chinese concepts by controlling the properties of the Equivalent class of classes and subclasses and the attributes of individual of the Same Equivalent class As.The Semantic Retrieval and Reasoning of Mongolian Medical ScienceThis paper uses Jena inference engine to realize diagnosis reasoning and prescription recommendation.It is a groundbreaking study of knowledge discovery in the field of Mongolian medicine through the use of semantic reasoning to judge symptoms and choice prescription.In clinical practice,it can play an important role through auxiliary diagnosis and misdiagnosis prompting.(3)Knowledge Base of Mongolian MedicineOn the basis of the ontology of Mongolian medicine,this article aims at the actual demands of Mongolian medicine clinical diagnosis,education,teaching and scientific researches,and develops the knowledge base in the area of Mongolian medicine.The knowledge base will cover all types of semantics in the Mongolian medicine field,with many functions such as semantic retrieval and reasoning,knowledge visualization,diagnosis aids,misdiagnosis prompts.Finally,it will achieve the transformation from the domain literature,domain information to domain knowledge.
Keywords/Search Tags:Traditional Mongolian Medicine, Ontology, Concept Lattice, Semantic Retrieval, Knowledge Discovery
PDF Full Text Request
Related items