Web Services Enabled Text Categorization System: Service Infrastructure Designing

Web services over the Internet have become a broadly used concept in a variety of applications for business, science, engineering and entertainment. To this end, this paper also proposes to create a new web service application, PSE-TC — a Problem Solving Environment (PSE) for text categorization that integrates the traditional text categorization system and the novel Grid technologies based on PSE infrastructure. In our previous paper (E-Science 2006), we have proposed the framework of PSE-TC which is able to support the activities that concern the building of the text classifier service, the classifying of the texts, the defining of the workflow, the selecting of service’s nodes and the reflection of the execution status through the web portal. In this paper, more web service modules are introduced in detail. Meanwhile, we have launched a novel middleware named Agent for (1) Executing those defined workflows by the users and locating and invoking the services on the Grid nodes according to those workflows, (2) saving the data produced by every service of the workflows, and (3) being a module as the liaison system uniforms the output data from preceding modules and/or external modules as the request of the next modules and transfer the data to the input data for the next module. By embedding this Agent service into our framework, a web service can be plugged into easily, and therefore the services modules connectivity and compatibility are improved to a great extent. In the end, some convincing experiment data are obtained, and prove that PSE-TC can provide a seamless and powerful text categorization research platform.