摘要
为了解决海量交通大数据实时预测问题,引入了Hadoop云平台结合K近邻非参数回归方法预测短时交通流。由于MapReduce框架的并行性,大大缩减了查找K个近邻的时间。通过实验证明,在集群上的预测时间相比在单机上的预测时间大大缩减。并且基于MapReduce框架的预测速度随着集群规模的增大而增大,表现出集群的可扩展性。该方法可以满足交通控制和交通诱导系统的实时性和精确性的需求。
In order to solve massive traffic big data in real- time prediction,we introduced themethod of combination of Hadoop platform and K nearest neighbor non- parametric regression to predict traffic flow. Because of the parallelism of MapReduce framework,it greatly reduced the time to find K nearest neighbors. It is demonstrated by experiments that the prediction time on the cluster compared with that on a single machine have been greatly reduced. And the forecasting speed based on MapReduce framework increased with the increasing cluster size,showing good scalability. This method can meet the needs of real- time and accuracy of Traffic Control and Traffic Guidance System.
出处
《无线通信技术》
2015年第3期38-43,共6页
Wireless Communication Technology
基金
国家"十一五"科技支撑计划资助项目(2006BAG01A0)
国家自然科学基金资助项目(10972027)
江苏大学校基金资助项目(11JDG064)
