摘要
飞行资料异常操作数据的有效准确挖掘是保证飞行安全,提高飞行训练水平的关键。提出一种基于双曲模型游散性牵引的飞行资料异常操作数据挖掘算法,针对多层异常操作数据在不同域之间的伪装变换特性,采用双曲非线性模型映射方法,对伪装域进行多层分离,依据访问数据游散性,对数据的深层次特征进行提取,以此为牵引,识别出异常操作数据。采用飞行资料的实际随机异常操作数据进行测试,结果显示,采用基于双曲模型游散性牵引的方法,异常操作数据的识别率达到了100%,在提高飞行安全等领域具有良好的应用价值。
The effective and accurate abnormal operation of flight information data mining is the key to ensure flight safety and the key to improve the level of flight training. An abnormal database access mining algorithm is proposed based on stray traction in hyperbolic model. The data characteristic between different domains is taken out. The hyperbolic nonlinear model mapping method is used. The multi-domain separation of camouflage is extracted out. The couples access to data based on in-depth feature data extraction is finished. Random abnormal operation data are used to do the effective abnormal mining. The test results show that with traction stray hyperbolic model. The identify rate of abnormal access data rates up to 100 %. It has good value for improving the flight security application.
出处
《控制工程》
CSCD
北大核心
2014年第4期617-620,共4页
Control Engineering of China
基金
国家自然科学基金项目(60875948)
