Date Added: Sep 2011
This paper introduces a new method to detect the Intrusion attacks on web services using XML similarity classifiers. First, pre-process methods are employed to distinguish the requests both based on the structure and content. WSDL description of the web service is employed to determine the valid structure of a request. A sequential threshold method is employed to define a request as an intrusion. Moreover, an optimized semantic similarity algorithm is employed to define the discrepancy between the contest of the request and the database of normal requests. Evaluation results are conducted based on various methods and algorithm which were presented in academia. The final section presents the promising way with the highest accuracy.