Study On Hierarchical Component Retrieval Based On Ontology

With the computers used widely and software requirements increasing, in order to industrialized software's production, software's reusability should arouse the concern of the field of software developing. The design and development of component repository also should is highly regarded for the increase of the reusable component. To use the existing components to develop the suitable software, enterprises need to solve the problem:how to search appropriate components from massive components. Components' effective retrieval is based on classification and representation. However, the traditional methods implement components retrieval just from grammar, without considering semantic information of component representation. Components description and classification based on ontology is probably to solve this problem.The users focus on component function information when they retrieve the right components, therefore, mining component function information should be research direction for scholars to improve component retrieval efficiency. Meanwhile, in order to make more reasonable component selection, users also should know the detail component matching information.In view of the above questions, the paper constructs component model using protege, and adopts W3C supplied OWL as component's description language. When designing retrieval system, this paper fully considers the semantic information of components' function, and then find components' hierarchical relationships through constructing function ontology algorithm based on FCA. Components retrieval results will show components matching degrees through the stratified retrieval strategy.Under the components description, the paper proposes ontology-based semantic component retrieval mechanism. The semantic retrieval method accomplishes a series of work as follows: calculate the semantic distance between vocabularies using wordnet model; get matching degree according to the component matching algorithms; give users the recommended component list according to the component function ontology.Retrieval efficiency is measured through recall and precision.The experiment results show that the method proposed in this paper can make the best of the semantic information contained in components, provide users sufficient component recommend information and improve components retrieval efficiency.