基于FP-Tree的QAR数据故障检测研究  

RESEARCH ON FP-TREE-BASED QAR DATA FAULT DETECTION

作　　者：杨慧[1] 赵兰草 

机构地区：[1]中国民航大学计算机科学与技术学院,天津300300

出　　处：《计算机应用与软件》2014年第10期41-44,共4页Computer Applications and Software

基　　金：国家自然科学基金与中国民航局联合项目(61179063/F01)

摘　　要：QAR(Quick Access Recorder)数据具有高维、复杂及数据量大的特性,严重影响数据处理效率。为降低其数据量与数据复杂性,高效检索并确定当前QAR数据是否是故障数据及其故障类型,首先通过PAA表示方法对QAR数据初步压缩,然后采用FP-Growth算法思想对压缩后的数据创建FP-Tree并只保留其频繁前缀子树,最后通过子树匹配确定测试数据与故障模型数据之间的匹配度。采用真实的飞机飞行QAR数据验证了算法的有效性和准确度。The characters of QAR （quick access recorder） data such as high dimensionality, complexity and large data volume make the data processing work seriously inefficient. In order to reduce its data volume and data complexity, as well as to retrieve efficiently and check out whether the current QAR data is the fault data and the type of the fault, we first compress QAR data initially through PAA representation method, and then adopt the idea of FP-Growth algorithm to create FP-trees on the compressed data but only keep their frequent prefix-trees. At last, we determine the matching degree between test data and fault model data using subtree matching. The effectiveness and precision of the algorithm are validated with real airplane flight QAR data.

关 键 词：频繁模式树 子树匹配 异常检测 

分 类 号：TP312[自动化与计算机技术—计算机软件与理论]