摘要
为了解决惯性领域内"数据丰富知识贫乏"的问题,建立一个基于数据挖掘的智能故障诊断系统,并重点围绕其中的数据挖掘环节展开研究。以全姿态组合陀螺TQZ-1A为研究对象,运用Clementine12.0工具,借鉴CRISP-DM行业标准,构建了基于两阶段聚类并做改进的C5.0模型。经过模型评价指标的综合评估验证了模型良好的预测性能,说明所建立的模型是科学的,适用于工程实践。
In order to solve the problem that data is rich but information is poor in the domain of inertia,this paper establishs in- telligent fault diagnosis system based on data mining and researching chiefly around the part of its data mining.Taking TQZ-IA as the research object and referring to CRISP-DM,it builds C5.0 model based on two-stage clustering and anomaly detect with Clementine12.0.At last,and validates the good predictive performance of the model by comprehensive assessment of the model evaluation,and proves that the model be scientific and can be adaptive for engineering practice.
出处
《工业控制计算机》
2012年第3期47-49,共3页
Industrial Control Computer
