Enhanced XML Encryption Using Classification Mining Technique for e-Banking Transactions

In this paper a novel approach is presented for securing financial Extensible Markup Language (XML) transactions using classification data mining (DM) algorithms. The authors' strategy defines the complete process of classifying XML transactions by using set of classification algorithms, classified XML documents processed at later stage using element-wise encryption. Classification algorithms were used to identify the XML transaction rules and factors in order to classify the message content fetching important elements within. The authors have implemented two classification algorithms to fetch the importance level value within each XML document. Classified content is processed using element-wise encryption for selected parts with “High” or “Medium” importance level values. Element-wise encryption is performed using AES symmetric encryption algorithm with different key sizes. An implementation has been conducted using data set fetched from e-banking service in one of the leading banks in Jordan to present system functionality and efficiency. Results from the authors' implementation presented an improvement in processing time encrypting XML documents.