Anomaly Data Flow Detection in Wireless Sensor Networks Based on Feature Completion
Due to missing or incorrect data perceived by nodes,abnormal data flow detection is hindered,resulting in issues with detec-tion accuracy,false positive rate,and energy consumption.Therefore,a feature completion based abnormal data flow detection method for wireless sensor networks is proposed.Missing and erroneous data in a specific environment are Estimated and completed based on the correlation between data flows within the sensor network.Date are extracted from the completed wireless sensor network data stream and data feature mining is completed to prepare for subsequent abnormal data flow detection.Support vector machines is used to separate normal and abnormal data,thereby achieving abnormal data flow detection in wireless sensor networks.The results show that after fea-ture completion,the accuracy of abnormal data detection in wireless sensor networks is maintained at over 99%,with a false alarm rate below 0.3%,and a maximum energy consumption reduction rate of 35.87%.The detection time is below 0.8 seconds.The proposed method has the advantages of high accuracy,low false alarm rate,low energy consumption,and short time using.
wireless sensor networkabnormal data detectionspatial correlationfeature miningsupport vector machine
国家科技期刊平台
NSTL
万方数据
