利用出租车GPS轨迹数据进行短时交通流量预测：以重庆市解放碑街区为例

交通流量预测是智能交通系统的重要组成部分。本文以重庆市解放碑街区为研究区域进行交通流量预测分析，基于研究区域内出租车GPS轨迹数据处理获取时间间隔为5min、10min、15min的交通流量序列。同时为充分挖掘交通流量序列特征规律，减小序列非线性、非平稳性带来的影响，本文提出一种基于信号分解的预测模型GE-RL。通过一般线性模型（GLM）将原始序列分解成周期序列、趋势序列和残差，同时引入经验模态分解方法（EMD）对残差进一步分解以充分挖掘序列特征；模型预测方面，构建随机森林模型（RF）对周期序列和趋势序列进行预测，接着引入长短期记忆网络模型（LSTM）构建RF-LSTM残差模型对EMD分解的各分量进行预测，通过叠加各模型预测成果得到最终预测结果；同时为验证模型精度，设置对照模型进行比对。结果表明，所构建的GE-RL模型在预测精度上均高于对照模型，可以满足基于不同样本时间间隔的交通流量预测的需要。

Short-term traffic flow forecasting using GPS track data of cabs -Take the Jiefangbei neighborhood of Chongqing city as example

Traffic flow prediction is an important part of intelligent transportation system. In this paper, traffic flow prediction analysis is carried out in the Jiefangbei neighborhood of Chongqing, and traffic flow sequences with time intervals of 5 min, 10 min and 15 min are obtained based on the processing of GPS trajectory data of cabs in the study area. At the same time, in order to fully explore the characteristic law of traffic flow series and reduce the impact of series nonlinearity and non-smoothness, this paper proposes a prediction model based on signal decomposition, GE-RL. The original series is decomposed into periodic series, trend series and residuals by the general linear model (GLM), and the residuals are further decomposed by the empirical mode decomposition method (EMD) to fully exploit the characteristics of the series. The final prediction results are obtained by superimposing the prediction results of each model; meanwhile, a control model is set to verify the accuracy of the model. The results show that the constructed GE-RL model is higher than the control model in terms of prediction accuracy and can meet the needs of traffic flow prediction based on different sample time intervals.