【摘要】 探究一种多出口流量数据高效清洗方法,能够提高流量数据高效清洗的同一性,降低出错率,增强多出口流量高效清洗的清洗效果,具备一定的实用性。针对当前方法在清洗多出口流量数据时,由于多个交换机设计结构不同导致高效清洗的流量数据与原流量数据存在不同一的问题,提出一种基于Rough集理论的多出口流量数据高效清洗方法,通过建立差分自回归移动平均模型,运用该模型对在某时间点含有残缺值的差分自回归移动平均模型进行计算,得到三种流量数据缺失值清洗模型,依据这三种模型得到扁平化网络多出口流量数据高效清洗模型。将清洗流量数据四元组作为出入,计算流量数据清洗函数值,输出高效清洗多出口流量数据,实现多出口流量数据高效清洗。仿真证明,所提方法能够减小噪声点的干扰,提高查全率,降低出错率,提高高效清洗的同一性。更多还原

【Abstract】 A method of efficient cleaning of multi-outlet flow data can improve the identity of efficient cleaning of data and reduce the error rate, which has a certain practicality. Therefore, a method of efficient cleaning for multi-outlet flow data based on Rough set theory was proposed. At first, our research established the differential auto-regression moving average model and used this model to calculate the differential auto-regression moving average model with incomplete value at some point, and then three kinds of traffic data missing value cleaning models were obtained. Moreover, this research obtained the efficient cleaning model for multi-outlet flow data in flattened network based on three models. Finally, based on the 4-tuple of cleansing flow data, the research calculated the flow data cleansing function value and outputted the multi-outlet flow data after the efficient cleaning. Thus, the multi-outlet flow data could be efficiently cleaned. Simulation proves that the proposed method can reduce the interference of noise point and improve the recall rate. Meanwhile, this method can reduce the error rate and improve the identity of efficient cleaning.更多还原