Abstract

With the rapid growth of Web services on the Internet, it becomes a great challenge for Web services discovery. Classifying Web services with similar functions is an effective method for service discovery and management. However, the functional description documents of Web services usually are short in their length, with sparse features and less information, which makes most topic models unable to model the short text well, consequently affecting the Web service classification. To solve this problem, a Web service classification method based on Wide & Bi-LSTM model is proposed in this paper. In this method, first, all the discrete features in the description documents of Web services are combined to perform the breadth prediction of Web service category by exploiting the wide learning model. Second, the word order and context information of the words in the description documents of Web services are mined by using the Bi-LSTM model to perform the depth prediction of the Web service category. Third, it uses the linear regression algorithm to integrate the breadth and depth prediction results of Web service categories as the final result of the service classification. Finally, compared with six Web service classification methods based on TF-IDF, LDA, WE-LDA, LSTM, Wide&Deep, and Bi-LSTM, respectively, the experimental results show that our approach achieves a better performance in the accuracy of Web service classification.

Full Text
Published version (Free)

Talk to us

Join us for a 30 min session where you can share your feedback and ask us any queries you have

Schedule a call