System uncertainty measure based on entropy and approximation classification quality in rough set

Uncertainty is an intrinsic feature of decision-making information systems and thus plays an important role in their performance. The system uncertainty can be measured efficiently by rough group theory based methods, which has better performance and certain drawbacks when combined with entropy. Based on all known entropy and approximation classification quality, we proposed a new measure method which avoids the deficiency of all known entropy measures effectively. The measures was verified reasonable and effective by realizing the independent learning and improved the overall performance of the Skowron algorithm.