基于小波分析的不确定数据流异常数据检测

针对数据流高速、无限连续和动态不确定性等特点,从提高不确定数据流数据管理能力的角度来解决不确定数据流中异常数据识别问题。首先采用小波分析,将连续数据流流量数据的高频与低频分量分离;其次,结合不确定数据流聚类方法找出数据中的异常点。仿真实验证明,该检测方法能够良好地适应数据流的不确定性,在一定条件下可获得相当好的检测效果。

An Anomaly Detection Method of Uncertain Data Streams Based on Wavelet Analysis

On account of data stream for high-speed,unlimited and dynamic characteristics of uncertainty,this paper proposed a solution to detect anomaly data,with respect of performance of management of uncertain data streams.By using the wavelet analysis to separate the high and low frequency components of continual data stream,then found anomaly data underlying in the uncertain data streams in combination with uncertainty clustering.Experimental results on actual data source show that this method can adapt to the uncertain of data streams well and provide precise instantaneous detection under certain conditions.