An Novel Artificial Immune System Approach to Robust Data Mining

We introduce several enhancements to deal with some of the weaknesses of previous artificial immune system models. Then, we present a framework for the accomplishment of several classical data mining tasks, such as frequent itemset discovery and robust clustering, based on ideas inspired from the natural immune system coupled with soft computing. For instance, we implement an artificial immune system mimicking the body’s adaptive learning and defense mechanism in the face of invading biological agents, to monitor and learn the user activities on a Web site. Like the natural immune system, the strongest advantage of immune based learning compared to current approaches is expected to be its ease of adaptation to the dynamic environment that characterizes several applications. We use both synthetic spatial data and real Web usage data to illustrate the workings of this novel computational paradigm.