Conceptual inductive learning: The case of unreliable teachers

摘要

Various algorithms for learning from examples usually suppose more or less reliable sources of information. In this paper, we study the influence of unreliable information sources on the learning process and the recovery possibilities. The problem is analyzed within the frame of the rough set theory which seems to be a suitable means for treating incomplete and uncertain knowledge. We briefly report on the learning system FLORA which is based on this theory.