基于深度学习的短时交通流量预测

Short-term Traffic Flow Forecast Based on Deep Learning

针对交通流数据的时间相关性和非线性等特点,现有预测方法未能充分获取交通流的本质特征,提出了一种基于深度学习的短时交通流量预测方法。该方法结合长短时记忆神经网络(LSTM)和支持向量机回归(SVR)作为预测模型,利用长短时记忆神经网络模型进行获取特征,用获取的特征训练支持向量回归进行交通流量的预测,比较了与其它模型的预测效果,真实数据集的结果表明,该模型有较高的预测精度。

According to the characteristics of time and non-linearity of traffic flow data,the existing forecasting method can not obtain the essential characteristics of traffic flow,and proposes a short-term traffic flow forecasting method based on depth learning.This method combined the short and long time memory neural network（LSTM）and support vector machine regression（SVR）as the prediction model.The long and short memory neural network model was used to obtain the characteristics,and the feature training to support the vector regression was used to predict the traffic flow.It was showed that the model has high prediction accuracy by the results of the real data set,and finally the prediction effect of the model with other models were compared.