基于基因表达式编程的话务量预测方法  

A Prediction Method of Telephone Traffic Based on Gene Expression Programming

作　　者：曾雪[1] 胡建华[1] 段磊[2] 

机构地区：[1]昆明理工大学信息工程与自动化学院,云南昆明650051 [2]四川大学计算机学院,四川成都610065

出　　处：《计算机仿真》2008年第7期170-173,共4页Computer Simulation

摘　　要：解决了在未知难以证明的先验知识时,传统时间序列分析方法无法得出精确的预测结果这一问题。将基因表达式编程(GEP)技术引入电信分析领域,提出了一种基于基因表达式编程的话务量预测模型GEP-Telecom。技术对某电信公司的话务量纪录进行的相关试验达到了预期的效果,给出了精确的预测。研究表明,GEP-Telecom模型在未知函数类型先验知识的情况下,能够自动进行属性筛选和属性融合,避免了传统预测方式的盲目性,提高了预测的精度。In this paper,to solve the problem that the traditional time series methods can＇t make an accurate prediction without the pre-knowledge,Gene Expression Program（GEP） is applied to the analysis of telecom portfolio and the GEP-Telecom model for telephone traffic prediction based on GEP is presented.By employing this technology,the experiment on telephone traffic records of X telecommunication reached the expectation and made a precise prediction.This research reveals that the technology can extract relative attributions and fuse them without the pre-knowledge about the function type,avoid blindness in traditional algorithms,and improve the precision of result.

关 键 词：数据挖掘 基因表达式编程 时间序列 话务量 

分 类 号：TP202.7[自动化与计算机技术—检测技术与自动化装置]