【摘要】 机场噪声中的异常情况拥有很大价值,利用它能够及时完善飞机和机场的设备。结合机场噪声数据的特点,对上述问题进行研究并提出一种基于改进的符号化聚集近似(Symbolic Aggregate Approximation,SAX)相似性度量的单监测点的时间序列异常检测方法。其运用相似性度量方法计算出度量结果,再运用k近邻异常检测方法进行异常发现,最后发现异常时间段。该方法在理论验证可行性之后在某机场的实测数据中进行应用,取得了良好的效果。更多还原

【Abstract】 With the expansion of airport transportation scale,the airport noise issue is becoming one of the obstacles for the sustainable development of the aviation industry. Anomalies in the airport noise are of great significance for the timely improvement of the equipments of aircraft and airports. In this paper,according to the characteristics of airport noise,a time series anomaly detection method for single monitoring point is proposed,which is based on the improved symbolic aggregate approximation similarity measurement. This method calculates the measure by applying the improved similarity measure,and finally finds anomalies using the k-nearest neighbor anomaly detection method. The proposed method is applied in practice after the theoretical verification of its feasibility.更多还原