关键词: RBD, MBD, 复合系统, 机器学习, 决策树

Abstract: With the gradual complication of some electronic system，using a single reasoning technology is more and more difficult to meet the demand of the fault diagnosis.Combining two or more reasoning technique is a trend of development of intelligent diagnosis，which can make use of their advantages and improve its efficiency.In this paper the intelligent diagnosis problem is discussed，a diagnosis architecture for the composite system is proposed，which combining rule-based diagnosis and model-based diagnosis.These two approaches work efficiently in different stage of the fault diagnosis system and with machine learning.

Key words: Rule-Based Diagnosis（RBD）, Model Based Diagnosis（MBD）, composite system, machine learning, decision trees