论文部分内容阅读
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 mining of flight data abnormal operation data is the key to ensure the flight safety and improve the flight training level. This paper proposes a data mining algorithm for flight data anomalous operation based on hyperbolic model of travel-diffusive traction. According to the camouflage transformation characteristics of multi-layer anomaly data in different domains, hyperbolic nonlinear model mapping method Layer separation, according to visit the data travel diffuse, the deep features of the data extracted as a traction, identify abnormal operation data. The actual random abnormal operation data of flight data were used to test. The results show that the recognition rate of abnormal operation data reaches 100% using the method based on hyperbolic model, which has good application value in improving flight safety and other fields.