Web Service Data Discovery Based on Contextual Feature Indexing And Timely Refreshing of crawled Threads

Web data service discovery has become a new sign of internet technology improvement for end users. User be- havior is usually sensitive to various contextual features which help in improving the semantic search performance significantly. However, contextual feature selection and extraction fails in providing result while discovering data in web services. The lack of precise user ratings on web as well as the sparse environment of data poses serious challenges to standard contextual feature extraction in terms of scalability and performance. The data discovery in the web does not create threads at a timely manner, and as a result, the user satisfaction degrades. This motivated us to address the problem by discovering the data based on the online leaning approach and include the feedbacks (i.e., ratings) from users. To improve the scalability rate of data discovery in web, Refresh Crawled Web based In- dexing (RCWI) mechanism is proposed in this paper. At first, dynamic web service data discovery is carried out through Semantic Source Root translator, In Semantic Source Root translator, the service requests (source) and responses is performed in one semantic root form for easy selection of contextual features. Secondly, Indexer is used to index data services for performing search and extrac- tion process. Then, Satisfied Data Discovery in RCWI mechanism is used to extract the most appropriate ser- vice based on user ratings. Finally, timing factor based thread creation helps to achieve user satisfaction and also perform the updates on refreshed threads. The data dis- covery process using RCWI improves the performance by applying neighboring cache mechanism which allows users to discover the services based on the online learn- ing experiences. Experiment is conducted on factors such as scalability rate, data discovery performance level, time taken on refreshing the crawled threads.Research scholar, Department of Computer Science, Karpagam University, Coimbatore, Tamil Nadu, India E-mail: a.r.rahuman@gmail.com, Associate Professor and Reader, Department of Computer Science, Periyar University, Salem, Tamil Nadu, India. E-mail: *ccsekar@gmail.com