【摘要】 针对当前方法监控大数据漏报率高、检测耗时长,导致防冲突检测效果差以及检测时效性差的问题,提出基于机器学习的监控大数据防冲突检测方法。通过计算监控大数据传输信道的占用率来估测信道负载情况,为提高信道负载估测的准确性,反复计算信道的占用率,检测监控大数据在信道传输过程中存在的冲突,利用数据包的传输时延来分析冲突,保证了监控大数据的优先传输;阐述监控大数据的冲突记录,并从客体和主体来划分记录的冲突信息;在此基础上,利用监控大数据中安全级别不同事件所发生的冲突时间计算冲突时间间隔,得到冲突间隔分布情况,并计算监控大数据的标准差,分析事件发生冲突的随机性以及规律性,通过检测监控大数据在信道传输过程中的冲突以及对冲突时间间隔的计算,最终实现了对监控大数据防冲突检测。实验结果表明,提出方法在对监控大数据防冲突检测时,数据的漏报率较低,检测效果和检测时效性较好。更多还原

【Abstract】 In the current method,the false negative rate is high,so this leads to poor detection effect and detection timeliness. Therefore,a method of anti-collision detection for monitoring big data based on machine learning was proposed. At first,the channel load was estimated by calculating the occupancy rate of big data transmission channel. In order to improve the accuracy of estimating the channel load,the occupancy rate of channel was calculated repeatedly. Meanwhile,the conflict of the monitoring big data during the channel transmission was detected and the transmission delay of data packet was used to analyze the conflict,so that the priority transmission of monitoring big data was ensured. Moreover,the conflict record of monitoring big data was explained,and the recorded conflict information was divided based on the object and the subject. On this basis,conflict time of events with different levels of security in the monitoring big data was used to calculate the conflict time interval,so as to obtain the conflict interval distribution and calculate the standard deviation of the monitoring big data. In addition,the randomness and regularity of event conflict was analyzed. By detecting the conflict of monitoring big data in the process of channel transmission and the calculating the conflict time interval,the anti-collision detection for monitoring big data was achieved. Experimental results show that the proposed method has low false negative rate and good detection effect and timeliness during the anti-collision detection of monitoring big data.更多还原