Design of Parallel Detection Method for Power Communication Data Traffic Anomaly
Data traffic anomaly monitoring is an indispensable part of power communication network operation,but the anomaly monitoring process is susceptible to the interference of different channel traffic redundancy,anomalous data types and other problems,which leads to a long time-consuming monitoring process and large errors.To solve the above problems,a parallel detection method for power communication data traffic anomaly is designed.Adaptive neighborhood algorithm is used to do dimensionality reduction of power communication data traffic,and the decomposition of power communication data traffic in different channels is realized by the idea of parallelism,which effectively reduces the redundancy interference of data traffic.The parallel detection method is designed to do parallel decomposition processing on the data traffic,and the decomposed data traffic is inputted into the detection model to complete the anomaly monitoring of the power communication data traffic by calculating the corresponding anomaly scores of the sample points.The experimental results show that the method has small mean square error,low time consumption and high recall rate.The performance for anomaly monitoring of power communication data traffic of this method is superior and has certain practical value.
Power communcationDimension reductionParallel decompositionRandom binary treeAbnormal scoreParallel detectionSmall errorRedundant interference
NETL
NSTL
万方数据
维普
