Component Facet Description And Retrieval Research Based On Component Credibility

With the development of computer technology,computer software has become an indispensable part of people's lives.Because of the particularity of the software,it cannot be the same as the factory production line.It is very important to improve software productivity and quality of software.Therefore,component-based software development is born.Through the component-based software development technology,it is recognized that the system assembling by selecting existing components can not only improve the work efficiency,but also reduce the development cost.However,the selection of the components to build software system blindly must cause component credibility problems inevitably.The larger the scale of the software system,the less trust that people will obtain.Credibility problem has become more and more serious.So that the selection of trusted components is the basis of building software systems.Nowadays,the reliability of components has become the research focus of component management and the research direction of component technology development.Since the domestic and foreign experts have made a lot of experimental research on the process of component production,management and reuse,there is little research on the credibility of component management and reuse phases.For the researches of component creditability,a unified model has not been established after analyzing the domestic and foreign literatures.Based on authoritative literature,the seven main characteristics and their sub characteristics of component credibility are analyzed and researched.Also,the metrics model of sub characteristics are proposed.The evaluation model of the component credibility is established.And the fuzzy comprehensive evaluation method is used to evaluate the component credibility.Then,the fuzzy theory can be used to solve the uncertainty and fuzziness in the process of component evaluation.On the basis of the existing classification description of the component,a new facet,the credibility facet,is added,which is helpful for the comprehensive description and accurate retrieval of the component.In the traditional component retrieval,each component in the component library is transformed into corresponding component description tree.At the same time,the retrieval condition that user inputs is transformed into the component query tree.At last,the above description tree and query tree of component are compared to match.The faceted retrieval method is proposed by combining tree matching and facet weight.Firstly,the credibility vectors of the query tree and the description tree are extracted.Then,a parts of high similarity components are selected from component library by using the method of cosine similarity calculation.Furthermore,the facet coding tree is set up for each component in the component collection after filtering the component collection.Finally,the string matching is used to retrieve the component second times.So a great deal of retrieval is avoided and the efficiency of component retrieval is improved.And in the calculation of matching degree,the weight of component facets is taken into account to improve the scientific and effective nature of component retrieval.According to the use of faceted description and retrieval method,a component library experiment system is established.Analyzing the recall and precision under the query conditions of different components,the feasibility and correctness of facet retrieval method combine tree match and the facet weight is verified.