The Research Of Knowledge Acquisition Method In Incomplete Information System

Knowledge acquisition on uncertain environment is one of key problems in Intelligence Information Processing. Many academician pay attention to how to acquire knowledge based on existed decision data in decision system when information is incomplete, imprecise or fuzzy. Recently, Rough Sets theory has been one of the important techniques; the polish scientist Z.pawlak brought forward this theory firstly in 20 century 80 years. It is a kind of tool for dealing with imprecise and uncertain problem. Evidence theory is.a kind of important reasoning method, Dempster first puts it forward in 1960's, it wants to use probability upper and lower limit to express the uncertainty in practical problem. Shafer make it a farther develop, make it systematize and theorize, form a kind of uncertainty reasoning theory-D-S evidence theory. D-S theory is often used in info- amalgamation and data processing.This thesis deals mainly with the knowledge acquisition methods as rough-sets, D-S evidence theory in incomplete information system, and obtains some significative result.This thesis discusses how to acquire knowledge in incomplete information system by Bayes theory, D-S evidence theory, analyse the relation and difference of them, and illuminate it by example.With respect to incomplete decision table, the thesis puts forward a kind of reduct algorithm based on the significance of attribute. The theses define the significance of attribute based on the dependence of attribute, act it as heuristic information, and picture the structure of algorithm,realize it by Visual Basic. The validity of this reduct algorithm is proved by test system.The work of this paper on the application of rough set and evidence theory in incomplete information system is valuable. But knowledge acquisition and incomplete information system is in a booming stage and there are many problems worth studying on the application of Rough Set Theory and D-S Evidence Theory in this field. Our work is just a beginning, and related work needs to be further developed.