Automatic Mission-Critical Data Discovery Based On Content: Experimental Results

In this work, we present results from a system that was designed to automatically identify critical data based upon content. This identification data can then be used for computer security purposes: known critical data can have extra safeguards extended to it and users that are attempting to access critical data illicitly can also be better identified. The process works by using a SVM-based filtering system that has been trained with expert knowledge. The system itself was described in a previous work; the contribution of this work is in the many experimental results which are presented. The results show that the proposed system will be a valuable tool in the computer security engineer’s toolkit.