Methods of learning rules based on rough set: LBR and LEM3

With the help of rough set theory, this paper puts forward a new way of machine learning - LBR (Learning By Rough set theory) - and then compares it with the algorithm LEM1 (Learning from Examples Method 1) that was proposed in "Incomplete Information Rough Set Analysis", Physica-Verlag Heidelberg, Ewa Orlowska (Ed.), 1998. From the comparison results, we find a new method of learning rules from examples, named LEM3, which is more flexible than LEM1. LBR and LEM3 have extensive application prospects in artificial intelligence.