Research Of The Software Components Retrieval Technology Based On Text And Graphic Processing

Component based software development is the precondition forlarge-scale software development, which combine with the specificcircumstance of required system, fully use ripe software products, willgreatly improve the efficiency of software development, and reduce a lotof duplication works, especially for the development of many applicationsin most industries MIS system, as well as cloud computing environment.With a large number of optional reusable software component is thefoundation of CBSD, and effective component retrieval mechanism canreduce the cost of software component searching and understanding. Thetraditional software component retrieval methods are mainly based onkeyword, facet and signature, due to the lack of domain information andthe differences of user semantic description, these methods are bad inprecision and recall, and require the user to have knowledge of thesoftware component library and retrieval mechanism. To resolve theseproblems, many domestic and foreign scholars and researchers haveproposed a semantics-based software component retrieval method, which canfully take the semantics and knowledge in the field of user queries intoaccount, however, the differences between the user understanding of theproblem domain, the natural language description that user familiar with,and the software component specification semantics, have increasing thecost of the user to retrieve.This paper presents a software component retrieval methods based ontext and graphic processing. On the one hand, restricted natural languagefor user to retrieve is allowed, calculation the similarity between userquery text and software component description text in software componentlibrary, return the pre-selected software component information byICTCLAS(Institute of Computing Technology, Chinese Lexical AnalysisSystem) technical and VSM (Vector Space Model), then user choose theappropriate software component; On the other hand, user can retrievesoftware component by the graphic and chart document(such as data flowdiagram, use case diagram, etc.). These graphic and chart document is the various models diagram from the software analysis and design stage, isinherent information and convenient to access and with a huge amount ofinformation, but also has a strong intuitive and is helpful for users’applicability judgment. This paper choose the software component dataflow diagram as example, by restricted the way of user drawing data flowdiagram, return pre-selected software component based on the elementscomparison, in and out calculation, then users choose the appropriatesoftware component.Each step of the software component retrieval based on text and graphicprocessing is reaserched, and realize the whole software componentretrieval system. Finally, the user query instances are demonstrated, inorder to verify the feasibility and effectiveness of the softwarecomponent retrieval method proposed in this paper, and this paper provideanother idea for the software component selection and evaluation.