线性回归在世博会散客流量预测中的实验

世界博览会在中国上海召开以来,认真做好每日入会人数的预测工作对提高世博会的服务质量具有重要意义。每日散客客流量的影响因素受到天气预期、接受信息量、周末因素、指定日因素、团体客流等变量的影响。运用多元线性回归对日散客客流量进行偏相关分析,通过对观察期间内的回归值、观察期间后两周的预测值与真实值的比较,发现对构建的多元线性回归模型需要进行扩充,加入其他预测元素进行修正可使模型的预测精度提升。同时,随着时间的推移,影响每日散客客流量的因素不断增加,存在众多不可解释的随机因素,需要不断修正该模型,不断提高模型的预测精度,做好每日入会人数预测工作,有助于资源的优化配置工作。

Implication of Multi-Linear Regression in Forecasting FIT Flow of World Expo

With World Expo opening in Shanghai,it is important for Chinese government to forecast the number of daily admission to improve the quality of service.FIT daily flow is influenced by weather expected,the amount of information received,the weekend factor,day factor designated,passenger groups and other variables.Through multiple linear regression and partial correlation analysis of FIT flow,and comparison of regression value between the observation period and the observation period with the forecast value and true value of two weeks,it was found that the multiple linear regression models need to be expanded to include other forecast elements modified to improve forecast accuracy.Meanwhile,with time passing by,the factors affecting the daily FIT flow were continuously increased and many inexplicable random factors came into existence,for which the model needs to be modified.In short,to continuously improve the forecasting accuracy,it is necessary to forecast the number of daily admission so as to help optimize resource allocation.