ARIMA模型在移动话务预测中的应用

电信运营商通过分析各个时间段、各个具体区域、热点区域的历史话务数据，能够对未来一段时间的话务量进行预测，从而提供面向运营商管理层的决策支持。因此，采用合适的方法对话务量进行准确预测，无论对于话务短期均衡，还是对于网络规划都有一定的帮助。本文将短期预测的代表方法——自回归求和滑动平均（Autoregressive Integrated Moving Average Model，简称ARIMA）模型应用于话务分析，并与往常使用的指数拟合趋势线方法进行对比，实验结果表明ARIMA预测精度较高。

The Application of the ARIMA Model in the Telephone Traffic Analysis

After having analyzed telephone traffic statistic,telecommunication operator can predict the traffic in future,in order to provide support for administration decision.Thus,to construct an appropriate model is not only useful for keeping balance of telephone traffic but also network planning.In this thesis,ARIMA model,a representative technique for short-term prediction is applied in traffic analysis.Compared to exponent,ARIMA shows a better performance in prediction precision.