Immunocomputing models computer systems after a body's natural immune system. Like a body's natural immune system, it attempts to detect, isolate, and remove foreign material or hacking attempts. Stide, an immuno-computing process model based on table lookup, has detected common intrusions in both artificial and live data in prior research. In our research, we investigated the value of using a more powerful process modeling technique. This process modeling technique, called fuzzy k-modes, clustered categorical patterns of system calls into centroids and memberships. These centroids and memberships then classified new process patterns as normal or abnormal. We obtained process patterns from an established data set for which stide results are known. Results for our data model were mixed. While acquiring the results, we established novel innovations aiding fuzzy k-modes. These novel innovations include a new index to test for data uniformity, a new logarithmic dissimilarity measure, a reduction in time complexity, and the failure to convert two well-known quantitative validity indexes to qualitative data.

Pizza and soda will be served for those attending!