Application of Improved Algorithm of Data Reduction to Knowledge Discovery of Information Security Management

Rough set theory has been a powerful methodology in data mining and knowledge discovery, extracting and minimizing rules from decision tables. There are mainly two kinds of ways for knowledge discovery: the one is to get specialized knowledge from experts in this fields, the second is to provide automated analysis solutions from database. But there are few studies that focus on the knowledge discovery combing specialized knowledge with automatic knowledge analysis. In this paper, rough set methodology is extended with a heuristic research algorithm. This algorithm, based on the discernibility matrices, integrates the frequency and significance of the attributes and the contribution rate of the rules to subjective judgment. This algorithm can find out the attributes with relative high subjective values. It is especially of importance to controllable system, where the value can be affected by the subjective judgment. And this algorithm is applied in the empirical studies in information security management.