Natural Language Processing Techniques for the Extraction of Semantic Information in Web Services

With the increase in the number of Web services and the value they are taking, discovery of services that meet the criteria of users becomes a major challenge. Currently the discovery process lacks of well-defined semantics, therefore it is difficult to obtain the non explicit meaning of services, such as functionality. In order to provide an accurate solution to the problem described above, in this paper we evaluate two natural language processing (NLP) techniques to extract semantic information from textual descriptions of Web services: linguistic patterns and extraction rules. Both techniques are implemented and compared in order to select the best alternative to our problem. This information will be useful for enhancing semantically the discovery process of services. We implemented a simulation environment for experimentation, and designed a set of experiments to show the applicability of our solution approach. Results of the tests show that the linguistic patterns are better than extraction rules.