基于混合学习算法IHMCAP的故障诊断模型

A FAULT DIAGNOSIS MODEL BASED ON HYBRID LEARNING ALGORITHM IHMCAP

在故障集和差错属性集的基础上 ,通过结合了基于概率论的符号学习与神经网络学习的增量式混合型多概念获取算法 IHMCAP寻找属性值与故障类型之间的对应关系 ,由此建立一个故障诊断模型 .实验表明 ,该模型不仅精度高、速度快、学习能力强 。

In this paper,a fault diagnosis model that uses an incremental hybrid multi concept acquisition algorithm IHMCAP is proposed based upon fault set and defective attribute set.The model combines probabilistic based symbolic learning and neural learning to search for the relationships between attribute values and fault types.Experiment results show that this fault diagnosis model not only achieves high accuracy,fast speed,strong learning ability,but also well balences the utility of domain knowledge and fresh data.