Knowledge Acquisition Method By "Rough Set" In The Expert System

Expert system is a kind of intellectual programming system can solve problems in relevant field with expert's levels, which can use the experience and special knowledge that domain expert accumulated for many years , simulate the thought process of human expert , solve the difficult problem that generally only expert can do . Knowledge acquisition is the key and " bottleneck "question in building expert system . In order to solve this problem, this thesis studies knowledge acequisition.Theory of Rough Set is a softly calculating method based on set theory, which recognizes that knowledge is a categorised ability that mankind and other species inherented and extend the concept of knowledge reduce relative reduce . dependence of knowledge. We can obtain decision rules through relative reduce which has removed redundant of attribute. Because it is unavoidable to introduce noise data and with the demand of rule expanding , this thesis adopts variable precision rough set mode approximate reduce was introduce from this model.While putting Rough Set theory into practice, this thesis pays attention to setting-up the proper data structure.In order to improve the data utilization ratio and promote rule quality, this thesis puts forward the method of "divide equally and examine each other This thesis bring forward the method of dynamic reduce to overcome data noise and confirm the best reduction Finally with the help of Rosetta tool software we apply the above concept and method to reality, and succeeded in obtaining the optimum rule for the expert system of production scheduling in DaYe Iron Ore Mine,WuHan Iron and Steel Company.