基于马尔可夫时变模型的流量数据挖掘

Traffic Data Mining Based on Markov Time Varying Model

随着网络技术的快速发展,目前网络的规模很大且有较大的复杂性,因此网络管理变得越来越困难和复杂,因此流量预测在网络管理中发挥越来越重要的作用。针对实际网络中收集到的大量实际流量数据,提出了一种基于时变网络的自适应网络流量预测算法,采用自适应学习率法,根据总误差增减变化趋势以及不同的改变来调整适应率;然后根据正向和反向的计算来校正各层的权重。仿真结果表明,与传统的时变网络相比,基于传播时变网络的自适应流量预测算法在预测结果中具有更好的性能,并具有较小的误差。

With the rapid development of network technology, the network management becomes more and more difficult and complex. The traffic prediction plays an increasingly important role in network management. For a large number of actual traffic data collected from the actual network, this paper presents a prediction algorithm based on time-varying network adaptive network traffic. According to the change trend of total error and adjust to different change rate, the algorithm is an adaptive learning rate method with the calculation of the forward and reverse to correct the weight of each layer. The simulation results show that compared with the traditional time varying network, the adaptive flow prediction algorithm based on the propagation time-varying network has better performance and smaller error in the prediction results.