一种短信数据量预测方法的研究

Research on a Predicting Method of Short Message Data Quantity

评估网络处理能力是否满足短信增加量,关键之处是及时、准确完成短信数据量的预测.本文根据短信数据量时间序列的非线性性质,给出了一种预测方法.这种方法基于相空间重构,应用混沌时间序列的加权一阶局域法单步预测模型预测短信数据量.以全国2004年1月到2007年1月的月短信数据量为实例进行计算分析,通过Matlab仿真,得出相对误差在0.017-0.08以内.

The timely and accurate prediction of short message data is a key to estimate whether the capability of a network affords increasing short message data. A predicting method of short message data is presented using the nonlinear characteristics of data. The method bases on the theory of phase-space reconstruction that adopts the add-weighted one-rank local-region forecasting model of chaotic time series. The calculation and analysis depend on the monthly short message data of China from Jan. 2004 to Jan. 2007 using the Matlab. The relative error of the predicting data is between 0. 017 and 0.08.