The Study Of The Query And Identification System Of Insect Based Bintree And Natural Language

As an application research of computer-aided decision technology in entomology, The query and identification expert system of insect, which is based on database technology and expert systems, has been developed.Users can identify species of forest insect of YUNNAN and aided by the query and identification expert system of insect based on bintree and natural language. The database contains 14 orders, 195 families of insects in YUNNAN forests and 5333 species of insects' Chinese name, latin name, character, geographical distribution, host and some picture.The system uses the relational database SQL Server2000 to construction two database about the insect identification and natural language repository, the database link to insects visit, inquiry, the identification module. The system provides multiple ways to identify and query information, such as browsing information, accurate query, ambiguous query, guidance query and natural language' s query, in order to meet different user requirements.The paper introduction a new design technology which based on Object-Oriented and user function. The expert system of YUNNAN forest insect identification is designed with PowerBuilder 10. 0. the system' s interface is designed like the style of the windows resource manager, in this way, the difficulties caused by the user's unfamiliaring with the system are reduced and the interface becomes concise, beautiful,human, easy-master and fine adaptability.The peculiarities of the system are as following:(1) the system is greater flexibility and feasibility, because users can identify the insect by several items. The range and velocity of identification is improved to overcome the limitation of searches of two items.(2) The paper has analysed the characteristics and the difficulty of insect species search, combining relational database to bintree theory, The system bring forward the coding principles of bintree linking number, and designed the reasoning mechanisms of identification module. The coding principles of bintree linking number overcome the problem about early heavy workload, long development period, Which consist in existing insect classification expert systems. The coding principles is an innovation.(3) The paper aim at the problem of existing expert systems of insects classification, which identification ways singleness, many steps to finish identification. In order to establish a close and friendly relations between human and computer, the paper seek after application natural language to the query of the insects characteristics. The paper analyse and compare the characteristics of many natural language processing technology, and the characteristics of insect taxology. The paper bring forward the natural language processing technology which combining limited natural language, the largest matching and keyword pattern matching law. Because the method need least language source, realization of the system is easer. The method is also an innovation of the system.(4) The query and identification of insect in the system is more exact and effective.The design method of the system have referenced worth to thj similar system.