基于Fisher的加权马尔可夫的话务预测

基于话务量数据的特性,提出基于加权马尔可夫的移动话务量预测模型。该模型具有要求的样本数量少,运算速度快,预测精度高,可检验等的特点。根据该模型的缺陷,应用模糊Fisher准则的有序聚类方法,可以对话务序列进行分类并建立分类标准,该方法克服了聚类有效性对样本空间分布的依赖并同时提高了算法的效率。仿真结果表明,在移动通信话务量预测中,该算法与其他方法相比较,运算速度快,准确度高。

A Fuzzy Fisher based Algorithm-Weighted Markov Chain for Forecasting Traffic Load

For the speciality of traffic load,a traffic load forecasting model based on weighted markov chain is proposed.This model requires less samples,operates rapidly and makes,forecast precisely,and is verifiable.Besides,this model,by applying the clustering algorithm of Fuzzy Fisher Criterion,classifies the traffic sequence and establishes classification standard.The proposed method could avoid the dependence on sample space distribution of the clustering validity,and raise the clustering efficiency.The experiment results indicate that the algorithm is more rapid and accurate as compared with other methods in mobile traffic load forecasting.