Automated knowledge acquisition is an important research issue in machine learning. There have been proposed several methods of inductive learning, such as ID3 family and AQ family. These methods are applied to discover meaningful knowledge from large database, and their usefulness is in some aspects ensured. However, in most of the cases, their methods are of deterministic nature, and the reliability of acquired knowledge is not evaluated statistically, which makes these methods ineffective when applied to the domain of essentially probabilistic nature, such as medical one. Extending concepts of rough set theory to probabilistic domain, we introduce a new approach to knowledge acquistion, which induces probabilistic rules based on rough set theory(PRIMEROSE) and develop an program that extracts rules for an expert system from clinical database, based on this method. The results show that the derived rules almost correspond to those of the medical experts.