Traffic forecasting has been an active research field in recent decades, and with the development of deeplearning technologies, researchers are trying to utilize deep learning to achieve tremendous improvements in tra...Traffic forecasting has been an active research field in recent decades, and with the development of deeplearning technologies, researchers are trying to utilize deep learning to achieve tremendous improvements in traffic forecasting, as it has been seen in other research areas, such as speech recognition and image classification. In this study, we summarize recent works in which deep-learning methods were applied for geospatial data-based traffic forecasting problems. Based on the insights from previous works, we further propose a deep-learning framework,which transforms geospatial data to images, and then utilizes the state-of-the-art deep-learning methodologies such as Convolutional Neural Network(CNN) and residual networks. To demonstrate the simplicity and effectiveness of our framework, we present a formulation of the New York taxi pick-up/drop-off forecasting problem, and show that our framework significantly outperforms traditional methods, including Historical Average(HA) and AutoRegressive Integrated Moving Average(ARIMA).展开更多
针对传统数据处理组合方法(Group method of data handling,GMDH)网络建模用最小二乘法辨识参数会导致模型预测效果不理想的问题,将模糊推理模型引入GMDH网络,以取代传统GMDH网络的部分描述(即完全二元二次多项式),提出了一种基于模糊G...针对传统数据处理组合方法(Group method of data handling,GMDH)网络建模用最小二乘法辨识参数会导致模型预测效果不理想的问题,将模糊推理模型引入GMDH网络,以取代传统GMDH网络的部分描述(即完全二元二次多项式),提出了一种基于模糊GMDH网络的交通流量预测模型。计算机仿真结果表明,该模型预测平均相对误差仅为2.31%,小于传统GMDH网络模型预测平均相对误差3.35%,说明了该模型是有效的。展开更多
基金funded by Shenzhen Municipal Development and Reform Commission,Shenzhen Engineering Laboratory for Data Science and Information Technology(No.SDRC [2015]1872)
文摘Traffic forecasting has been an active research field in recent decades, and with the development of deeplearning technologies, researchers are trying to utilize deep learning to achieve tremendous improvements in traffic forecasting, as it has been seen in other research areas, such as speech recognition and image classification. In this study, we summarize recent works in which deep-learning methods were applied for geospatial data-based traffic forecasting problems. Based on the insights from previous works, we further propose a deep-learning framework,which transforms geospatial data to images, and then utilizes the state-of-the-art deep-learning methodologies such as Convolutional Neural Network(CNN) and residual networks. To demonstrate the simplicity and effectiveness of our framework, we present a formulation of the New York taxi pick-up/drop-off forecasting problem, and show that our framework significantly outperforms traditional methods, including Historical Average(HA) and AutoRegressive Integrated Moving Average(ARIMA).
文摘针对传统数据处理组合方法(Group method of data handling,GMDH)网络建模用最小二乘法辨识参数会导致模型预测效果不理想的问题,将模糊推理模型引入GMDH网络,以取代传统GMDH网络的部分描述(即完全二元二次多项式),提出了一种基于模糊GMDH网络的交通流量预测模型。计算机仿真结果表明,该模型预测平均相对误差仅为2.31%,小于传统GMDH网络模型预测平均相对误差3.35%,说明了该模型是有效的。
