The idea of Knowledge Discovery is used to solve the dificulty decision process knowledge acquisition. The concept of Process Decision Data Table (PDDT) is defined to store process data. An improved process decision tree learnning arithmetic is proposed,which is based on CLS. ID3 is used to measure the fundamentality of condition attribute to decision rules in PDDT. The principle of process decision tree predigestion and the method of translateing it to decision rules are given.
