基于随机波动条件的公交客运量预测模型

针对传统的灰色GM（1,1）预测模型在预测公交客运量中存在误差过大的问题,结合公交客运量随机波动的显著特征,通过对残差序列进行再处理,构造新的数据序列,构建GM（1,1）改进预测模型对公交客运量进行预测,并应用于某城市的2条公交线路客运量预测。结果表明：随机波动条件下的GM（1,1）改进预测模型,使用预测序列与残差序列绝对值之和来构造新序列,对新序列进行建模后预测的公交客运量的平均相对误差分别为4.9%和5.3%,明显优于传统GM（1,1）模型预测的公交客运量的平均相对误差7.5%和7.45%;相对误差最大值分别降低了4.68%和2.99%。

Prediction model of public transport passenger volume based on random fluctuations

As there is a large error in public traffic volume predicted by traditional GM(1,1) model,it is imperative to modify it based on the random fluctuation of public traffic volume by rearranging residual sequence which produces new sequence data.The new modified GM(1,1) model is employed in measuring public traffic volume in two bus routes in a city.The modified GM(1,1) model under the condition of random fluctuation uses the sum of absolute value of predicted sequence and residual sequence to construct a new sequence.The result shows that,with the new modified mode,the average relative predicting errors of the two routes are 4.9% and 5.3%,but the ones by the traditional GM(1,1) are 7.5% and 7.45%.The maximum of relative error decreases by 4.68% and 2.99% respectively for each route.