Research On Semi-Automatic Construction Of Ancient Agronomy Ontology And Its Semantic Retrieval

Chinese agriculture history literature research has changed to digital construction in recent 20 years. It has been strengthened by modern technology that the generation, organization, access, communication and usage of agriculture history information resource. The researcher of agriculture history can get research data more easily and quickly. So the digital construction could promote the development of agriculture history research. Digital construction of agriculture history resources has made great progress, but the organization and access technology is still antiquated. The reason is that computer treated the user's keywords only as simple symbols. So the useful method to increase the information service of agriculture history is that building a semantic description mechanism to make computers have semantic understanding ability.Ontology has been paid much attention since it was proposed, which was a concept mould tool in semantic and knowledge hierarchy description. Ontology has been broadly applied in knowledge engineering, digital library, software reuse, information retrieval and Web heterogeneity processing and Semantic Web.This paper introduced ontology into agriculture history field, tried to construct ancient agronomy ontology as tools of agriculture history information processing, organization and usage to resolve semantic heterogeneity, which can make agriculture history concept have clarity and unique definition, promote communication between person and machine.The Semantic Web relies heavily on the formal ontologies that structure underlying data for the purpose of comprehensive and transportable machine understanding. Therefore, the success of the Semantic Web depends strongly on the proliferation of ontologies, which requires fast and easy engineering of ontologies and avoidance of a knowledge acquisition bottleneck. Manual construction and description of domain-specific ontology is a complex and time-consuming process. The recent study on ontology design methodologies shows that it is very hard for a designer to develop accurate and consistent ontology .Therefore, this paper emphasized on the research of semi-automatic construction ancient agronomy ontology and its semantic retrieval mechniasm.The main content of this paper can be abstracted as following:(1)Survey and Analysis of model of Information Organization of Agriculture History Information ResourcesWith the rapid development of computer technology and network technology, Internet is gradually becoming an important way to obtain information resource for the researcher. History researcher discarded the low effect information acquire way which is to search data from a huge of traditional physical literature. Digital information resource saved the time of searching literature of reasearchers. This paper surveys on the state of information organization of agriculture history information resource. The main mode of its information organization included of professional website, professional database and digital library/museum.Digital construction of agriculture history information resource is to supply better information service for the agriculture history researcher. It is the key to the digital construction of agriculture history information resource that whether can give better effect of agriculture history information services. This paper analyzed the state of agriculture history information organization based on the survey of agriculture history information resource, through the three layer of information organization mode, retrieval technology and information service type. By the analysis, we find that the shortage in the current information organization is that the system is lacking of semantic control mechanism. Keywords in the resource organization only be treated as symbols, not considering its semantic meaning.An ontology is an explicit specification of a conceptualization. The term is borrowed from philosophy, where an Ontology is a systematic account of Existence. By defining shared and common theories, ontology helps both people and machines to communicate concisely, supporting the exchange of semantics and not only syntax.(2) Research on the Pattern of Ancient Agronomy Ontology ConstructionThough ontology engineering tools have become mature over the last decade, the manual Acquisition of ontologies still remains a tedious, cumbersome task resulting easily in a knowledge acquisition bottleneck. In fact, these problems on time, difficulty and confidence that we ended up with were similar to what knowledge engineers had dealt with over the last two decades when they elaborated on methodologies for knowledge acquisition or workbenches for defining knowledge bases. A method that proved extremely beneficial for the knowledge acquisition task was the integration of knowledge acquisition with machine learning techniques. The drawback of these approaches, e.g. the work described in, however, was their rather strong focus on structured knowledge or data bases, from which they induced their rules.Therefore, a number of approaches propose to improve ontology construction using automatic discovery of taxonomic and non-taxonomic relationships from domain data or domain-specific texts. Unfortunately, in the approaches available, there is a lack of combination of the two methods, because methods for learning ontological relationships rely to a given initial taxonomy of concepts and use it in learning process. So, in this reseach, we take the approach that constructed the ontology semi-automatically. The domain expert gives the skeletons of ontology based on his background knowledge and current classification schema and theasrus. Then automatically acquired concepts relation by integrating the large scale statistic method and natural language processing method to expend and update the skeletons.③Research on Semi-Automatic Construction Technology of Ancient Agronomy OntologyThis paper attempted to take a method that extremely beneficial for the knowledge acquisition task was the integration of knowledge acquisition with machine learning techniques to increase the ontology construction effect. In the Construction, this paper integrated many methods into the recognition and identified of domain relation. This paper basically realized the function of domain concepts acquisition, taxonomy relation recognition, non-taxonomy relation recognition and ontology formalization description.①Automatic Acquisition of Domain Concepts of Ancient Agronomy OntologyThis paper adopted an approach of Non-dictionary Chinese word Segmentation techniques based on N-Gram to acquire domain candidate concepts. Then it took the approach to select core concepts from the candidate concepts, which adopted the automatic subject indexing method to get the core concepts based on the principle of literature guarantee.②Recognition of Taxonomy Relation of Ancient Agronomy OntologyFirst get the skeleton of ancient agronomy ontology based on the knowledge background of domain expert and the classification schema of current classification and thesaurus related to the ancient agronomy. It can assure that the constructed ontology has the guarantee of better general character and well logic foundation. Also the paper adopted the improved Agglomerative Hierarchical Clustering algorithm to recognition taxonomy relation from ancient agronomy corpus, which can expand and update the skeleton of ancient agronomy ontology acquired from domain expert and current classification schema and thesaurus.③Recognition of Non-Taxonomy Relation of Ancient Agronomy OntologyThis paper adopted the approach of integration the method based of Association Rule Mining and the method based of Natural Language Processing into the recognition of domain concept property relation from the ancient agronomy research corpus. The research Used the parameters of confidence and support to acquire the most associated concepts from the corpus, also following the characteristics of Chinese language syntax, we extracted subject, predicate and object of sentences. This triangle data can be treated as the triplet of Data Type and Object Type Property. This combination method can decrease the shortcoming of large scale statistical method which lacking necessary semantic logic foundation, the method can also avoid deficiency of the semantic relation analysis of concepts which excessively depended on complicated language processing model, as we know the model can't be acquired easily. In this research, we take the method to acquire synonym property for the concepts based on the mode matching.④Fomaliztion Description of OntologyOntology formalization file can be created automatically by batch processing according to decided rule and concept relation. This approach can increase the effect of ontology creation. This formalization was based on the foundation of recognized concept relation.(4)Research on the Semantic Retrieval Mechanism based on Ancient Agronomy OntologyThis paper designed and developed a semantic retrieval prototype based on ancient agronomy ontology in order to probe the mechanism of semantic retrieval. The prototype was consisted of retrieval words analysis module, semantic reasoned module, ontology browsing module and semantic query module. By the ontology, the prototype can get implied concepts of users through the semantic analysis of user's keywords. Then the machine can acquire the unified understanding to users; form the standardization description of concepts. The prototype realized the semantic retrieval based on the domain ontology. It proved that the semantic retrieval can get better results than the keywords retrieval through the contract test by the keywords retrieval. It also proved that the semantic retrieval can increase the retrieval effect. It takes challenges and chances to agriculture heritage information by the high development of information technology. How to supply better agriculture heritage information services is the crucial mission to the information services institution. This paper led Ontology technology to the agriculture history information management. It takes a punt at agriculture ontology semi-construction for agriculture history information management, which integrates information science, machine learning and natural language processing technology into a system for its semi-construction. This paper designed and developed a semi-construction system of agriculture history ontology, also included a semantic retrieval system based on the ontology. By the reason of time and person, this paper only selected a part of agriculture history—Ancient Agronomy as the research object. The system is only a Prototype which needs further research and a step forward revise.