摘要
大部分关联规则挖掘算法仅以出现频率作为阈值来产生关联规则,然而在发动机故障数据分析领域,很多关键故障的出现机率较低,采用传统的挖掘算法无法适应航空发动机故障分析的需要。提出一种新算法,称为权重式多重支持度关联规则挖掘算法,此算法可针对不同重要度的故障定出不同的支持度阈值,由此产生的关联规则,可以解决高重要度但出现较少的故障不易被挖掘的问题。
Most of the previous approaches of mining association rules only take frequency as thresholds to produce association rules. However, in the field of aero-engine failure analysis, those traditional approaches don＇t hold true at all, because of the low frequency of some important failures. The paper puts forward a new algorithm-weight algorithm with multiple supports for mining association rules, which can set different supportive thresholds in accordance with different significance to solve the problem that failure with high significance but low frequency is difficult to find out.
出处
《火力与指挥控制》
CSCD
北大核心
2009年第2期55-58,共4页
Fire Control & Command Control
关键词
数据挖掘
关联规则
权重式多重支持度
故障分析
data mining,association rules ,weight algorithm with multiple supports ,failure analysis
