摘要
针对城市轨道客流培育发展大致呈S形曲线的特点,首先以灰色Verhulst模型对西安地铁2号线客流进行预测.为提高预测结果的精度,对原始数据形成的序列数据通过对数变换处理来减少其波动性.由于所研究轨道线路尚未到达客流饱和阶段,为了规避单一模型较大的风险性,针对不同模型的特点,从处理过的原数据列选取西安地铁1号线开通年度(2014年1月)之后的西安地铁2号线数列,此数据列数据量较少,波动性小,采用灰色GM(1,1)模型对处理过的西安市地铁2号线数列进行预测.通过将灰色GM(1,1)模型预测结果与改进Verhulst模型预测结果进行线性组合,之后将采用不同预测模型的预测结果与实际值进行对比分析,发现组合模型的预测精度更高。
Aiming at the characteristics of urban rail passenger flow development,the S-shaped curve is mainly developed. Firstly,the gray Verhulst model is used to predict the passenger flow of Xi’an metro line 2. To improve the accuracy of the prediction results,the sequence data formed by the original data is processed by the logarithmic transformation to reduce the volatility. Since the studied track circuit was not yet reached the passenger flow saturation stage,in order to avoid the large risk of the single model,we select the Xi’an metro line 1 from the processed original data column for the opening of the year 2014 after the January 1st line of Xi’an metro line 2. This data column has the less data and less volatility. The gray GM(1,1)model is used to predict the number 2 series of Xi’an City. By linearly combining the grey GM(1,1)model prediction results with the improved Verhulst model prediction results, the prediction results of each model are compared with the actual values,and the combined model has a higher prediction accuracy.
作者
贾云蒲
陈宽民
曹夏玲
JIA Yunpu;CHEN Kuanmin;CAO Xialing(School of Highway,Chang’an University,Xi’an 710018,China)
出处
《河南科学》
2019年第5期840-846,共7页
Henan Science
基金
国家自然科学基金项目(71871027)
