A personalized search robot is developed as one major mechanism of a personalized software component retrieval system. This search robot automatically finds out the Web servers providing reusable software components, extracts needed software components from servers, classifies the extracted components, and finally establishes their indexing information for local component retrieval in the future. For adaptively tuning the performance of software component extraction and classification, an adaptive thesaurus and an adaptive classifier, realized by neuro-fuzzy models, are embedded in this search robot, and their learning algorithms are also developed. A prototype of the personalized software component retrieval system including the search robot has been implemented to confirm its validity and evaluate the performance. Furthermore, the framework of proposed personalized search robot could be extended to the search and classification of other kinds of Internet documents.
Analysis of lifecycle models and Software Component Retrieval
