针对电信行业自适应智能话务预测模型的分析与实现

《针对电信行业自适应智能话务预测模型的分析与实现》由会员上传分享，免费在线阅读，更多相关内容在学术论文-天天文库。

1、中文摘要法等。通过一系列模拟实验，说明该预测算法对日常话务量的预测效果比较好。(4)提出一种节假日话务量预测算法作者研究出一种节假日自适应话务预测算法，该算法采用了时间序列的水平和速度分析法、遗忘因子参数估计法、以及集合分组划分等方法。利用该预测算法对节假日进行话务预测的时候，它具有适应性强、关联程度高、预测效果好等优点。通过节假且话务量对比预测实验，说明该预测算法比较适合用于节假日的话务量预测。(5)提出预测算法软件化方法作者论述了预测方法软件化的基本原理，提出了预测软件工作的具体流程，并设计出了四川移动智能预测系统的基本结构，再根据所研究的自适应话务预测算法，设计和实现了话务预测软件

2、。目前，该话务预测系统已经在四川移动得到了运用，并取得了好的效果。关键词：话务预测，时间序列，最小二乘法，遗忘因子，预测算法ⅡAbstractAlongwiththeprosperityofChineseeconomy,thetelecommunicationindustryhasgotmanyextraordinaryachievementsinourcountry．Eachtelecommunicationcorporationhasaccumulatedlargeamountsoftrafficdata．rIhmcdataalsohasthenatureoftimeseries．S

3、oweCallusethebasictheoriesandthemethodsoftimeseriestoanalyzethesehistoricaltrafficdata，Thenfindouttherulesofmeirchanging,andestablishsomereasonabletrafficforecastmodelstoimplementauto-adaptedtrafficforecastfortheconditionsofcommondaysandfestivals．Accurateandreasonableforecastresultscallbeasthebas

4、icofdecision-makingforeachtelecommunicationcorporation’snetworkmaintenanceandmarketprogramming，etc，Thereforetrafficforecasthasgreatofsignificance．Thisdissertationmainlystudiesthetelecommunicationindustry’sauto—adaptedintelligencetrafficforecastmodalanditsrealization，itsmaincontentsinclude：basicth

5、eoriesoftimeseriesandcommonforecastmodels，thecycleanalysisoftraffictimeseries，trafficforecastmodelresearchfortheconditionofcomlTlondays，trafficforecastmodelresearchfortheconditionoffestivals，therealizationoftrafficforecastmodelandsoon．Thedissertationisacomprehensivesumming-uptoauthor’Sresearchwor

6、k～Itskeycontentsincludeasfollows：(1)ContrastingandanalyzingseveralkindsofcommontimeseriesforecastmodelsandtheirapplicationThecorrelationtheoriesoftimeseriesmodelingaretheacademicbasicwhichthisdissertationstudies．Forthequestionoftrafficforecastmodeling,theauthorhascontrastedandanalyzedseveralkinds

7、ofCommontimeseriesforecastmodels,andaccordingtothecommoncriterionsoftimeseriesmodels’choice，hasmadedetailedanalysisandexplanationsfortheirconcreteapplications．(2)Puttingforwardakindoftraffictimeseriescycleanalysisalgor