多媒体网络负面信息数据检测仿真研究
|
摘要
针对当前网络负面信息检测方法存在假阳性率和漏检率较高的问题,提出基于朴素贝叶斯的多媒体网络负面信息数据检测方法。通过FIR滤波法实现负面信息数据检测前的抗干扰滤波,对采集到的多媒体网络信息数据中心流量实行傅里叶变换操作,得到FIR滤波器输入向量。将输入向量设置为平稳向量输入至滤波器,计算滤波器传输函数,获取FIR级联滤波器。对滤波器的抽头系数进行计算,实现媒体网络干扰滤波。利用经FIR滤波器处理过的多媒体网络信息数据流量测试集,构建贝叶斯概率表达式,得到网络中待检测信息数据隶属某类别的概率。根据概率计算式获取负面信息数据检测目标函数,计算目标函数中的先验概率值,并使用先验概率值完成网络负面信息数据检测。仿真结果表明,所提方法可将检测结果的假阳性率控制在5%以下,漏检率约为0.67%。该方法检测率高且漏检率低,与当前相关方法相比具备较强的检测性能。
Due to the problems of false positive rate and high missed detection rate in current network negative information detection methods,this paper puts forward a method to detect multimedia network negative information data based on Naive Bayesis proposed. Firstly,the FIR filter method was used to achieve the anti-interference filtering before the negative information data detection,and then the Fourier transform was performed on the collected central flux of multimedia network information data to obtain the input vector of FIR filter. Secondly,the input vector was set as the stationary vector,which was inputted to the filter,and then the filter transfer function was calculated to obtain FIR cascade filter. Thirdly,tap coefficients of the filter were calculated to achieve interference filtering of the media network. Moreover,the test set of multimedia network information data traffic processed by FIR filter was used to construct Bayesian probability expression,so as to obtain the probability that the information data to be detected in the network belonged to a certain category. According to the probabilistic formula,the objective function of negative information data detection was obtained. Finally,the prior probability values in the objective functions were calculated.Thus,the network negative information data detection was completed by using the prior probability value. Simulation results show that the proposed method can control the false positive rate of detection result below 5%. The missed detection rate is about 0.67%. The method has high detection rate and low missed rate,which has stronger detection performance than current methods.
Due to the problems of false positive rate and high missed detection rate in current network negative information detection methods,this paper puts forward a method to detect multimedia network negative information data based on Naive Bayesis proposed. Firstly,the FIR filter method was used to achieve the anti-interference filtering before the negative information data detection,and then the Fourier transform was performed on the collected central flux of multimedia network information data to obtain the input vector of FIR filter. Secondly,the input vector was set as the stationary vector,which was inputted to the filter,and then the filter transfer function was calculated to obtain FIR cascade filter. Thirdly,tap coefficients of the filter were calculated to achieve interference filtering of the media network. Moreover,the test set of multimedia network information data traffic processed by FIR filter was used to construct Bayesian probability expression,so as to obtain the probability that the information data to be detected in the network belonged to a certain category. According to the probabilistic formula,the objective function of negative information data detection was obtained. Finally,the prior probability values in the objective functions were calculated.Thus,the network negative information data detection was completed by using the prior probability value. Simulation results show that the proposed method can control the false positive rate of detection result below 5%. The missed detection rate is about 0.67%. The method has high detection rate and low missed rate,which has stronger detection performance than current methods.
引文
[1]姜红红,等.基于大数据的电力信息网络流量异常检测机制[J].电信科学,2017,33(3):134-141.
[2]王影.无线网络流量异常数据信息检测仿真[J].计算机仿真,2017,34(9):408-411.
[3]侯爱华,高伟,汪霖.基于逻辑回归模型的流量异常检测方法研究[J].工程数学学报,2017,34(5):479-489.
[4]刘汝隽,贾斌,辛阳.基于信息增益特征选择的网络异常检测模型[J].计算机应用,2016,36(a02):49-53.
[5]郝亚洲,等.面向网络舆情数据的异常行为识别[J].计算机研究与发展,2016,53(3):611-620.
[6]赵新杰,等.基于迁移学习和D-S理论的网络异常检测[J].计算机应用研究,2016,33(4):1137-1140.
[7]黄伟,陈昊,郭雅娟.融合领域知识的网络异常检测方法[J].南京理工大学学报,2016,40(2):229-235.
[8]应斐昊,等.基于KTLAD的电力数据网业务流量异常检测[J].北京邮电大学学报,2017,40(S1):108-111.
[9]吴丽云,等.基于PLS特征提取的网络异常入侵检测CVM模型[J].控制与决策,2017,32(4):755-758.
[10]张峰,郑洪源,丁秋林.基于信任和权重的无线传感器网络数据融合模型[J].计算机科学,2017,44(5):37-41.
[2]王影.无线网络流量异常数据信息检测仿真[J].计算机仿真,2017,34(9):408-411.
[3]侯爱华,高伟,汪霖.基于逻辑回归模型的流量异常检测方法研究[J].工程数学学报,2017,34(5):479-489.
[4]刘汝隽,贾斌,辛阳.基于信息增益特征选择的网络异常检测模型[J].计算机应用,2016,36(a02):49-53.
[5]郝亚洲,等.面向网络舆情数据的异常行为识别[J].计算机研究与发展,2016,53(3):611-620.
[6]赵新杰,等.基于迁移学习和D-S理论的网络异常检测[J].计算机应用研究,2016,33(4):1137-1140.
[7]黄伟,陈昊,郭雅娟.融合领域知识的网络异常检测方法[J].南京理工大学学报,2016,40(2):229-235.
[8]应斐昊,等.基于KTLAD的电力数据网业务流量异常检测[J].北京邮电大学学报,2017,40(S1):108-111.
[9]吴丽云,等.基于PLS特征提取的网络异常入侵检测CVM模型[J].控制与决策,2017,32(4):755-758.
[10]张峰,郑洪源,丁秋林.基于信任和权重的无线传感器网络数据融合模型[J].计算机科学,2017,44(5):37-41.