摘要
鸟击问题严重威胁航空器运行安全,给航空业造成了巨大的经济损失,为了有效预防民航鸟击事件的发生,根据数据挖掘理论,在分析民航鸟击事件关键诱发属性基础上,提出了一种基于FPGrowth算法的民航鸟击事件关联性分析方法。根据中国民航鸟击事件统计数据,挖掘出鸟击事件各属性间潜在的、有价值的关联,通过设置最小支持度和最小置信度,得出重要的关联性规则。结果表明,该方法根据历史数据可推测出导致鸟击事件发生的相关因素,改善了以往凭借专家经验的片面性、模糊性和不确定性。通过飞机发动机设计、颜色涂装等措施切断导致鸟击事件发生的相关因素,达到有效预防鸟击事件的效果,完善防治措施,最大限度地避免鸟类撞击航空器,保障民航运输安全。
The given article is inclined to present a correlation analy- sis on the likely bird strike events against the civil aviation based on the FP - Growth algorithm. As a matter of fact, in recent years, bird strikes have gradually formed and turned to be a serious threat to the safety control and management of the civil aviation Night with the fast pace of our civil aviation development and the continuous improve- ment of the biological and ecological environment. In order to effec- tively prevent and reduce the incidence rate of such bird strikes, we would like to propose a FP - Growth algorithm with the help of data mining theory in hoping to present a more useful analysis method for the correlation of the civil aviation bird strike events and the likely dangers such those events that may bring harm to the civil aviation. As compared with the other similar algorithms, the FP - Growth algo- rithm we have chosen to adopt here enjoys the advantages of reducing the database scanning time and greater execution efficiency. What＇ s more, the bird strike transaction database containing the frequent item sets can be compressed v/a a frequent pattern tree by scanning the database twice. Besides, it is also possible to convert the problem of discovering the frequent patterns into the recursively short modes, which can significantly reduce the searching cost. Based on the re- search over the key involved attributes to such bird strike events, such as the aircraft types, colors, the flight phases and the flying time, the influential parts and so on, we can build up the bird strike FP-tree in accordance with the bird strike statistics provided by the China civil aviation office. In addition, we have also worked out all the potential and valuable associations of the attributes so as to formu- late the important association rules by setting up the minimum sup- porting coefficient and the minimum confidence coefficient promptly, by which it is possible to sum up the rules that the bird strikes can come out according to the numerous historical records of the past events. The results of our study indicate that the given method can help to infer the related factors leading to the bird strike according to historical data and overcome the lopsidedness and ambiguity with the help of the former expertise. Therefore, the security of civil aviation transportation and traffic can be guaranteed securely by enhancing the bird strike prevention capacity and evading unhappy events caused by the bird striking.
出处
《安全与环境学报》
CAS
CSCD
北大核心
2016年第1期110-114,共5页
Journal of Safety and Environment
基金
国家自然科学基金项目(71303110
71573122)
关键词
安全工程
鸟击
数据挖掘
FP-GROWTH算法
关联性
防治措施
safety engineering
bird strike
data mining
FP-Growth algorithm
correlation
prevention and： control ： 0 ,measures
