Research on short-term Traffic flow Prediction Based on Big Data Environment

In view of the rising index of traffic flow data, how to research and analyze the massive data quickly and accurately has become a hotspot in the field of Intelligent Transportation. This article conducted a study on the accuracy and instantaneity of the short-term traffic flow forecast in big data environment. A combined forecast model based on cluster analysis and Regression Forest in Spark platform was proposed. Through the application of Kmeans, it analyzed the temporal and weather’s correlation of traffic flow by cluster analysis method. In combination with the parallel Regression Forest algorithm as a predictor, Spark distributed computing cluster was deployed and massive data was combined for training and forecast. The experimental results show that the combined forecast model for equivalent data sets is more real-time than the stand-alone operation in the Spark cluster environment. And as the amount of data increases, the advantage is more obvious. Compared with stand-alone support vector machine and regression forest classic machine learning model, it also has better prediction results in terms of forecast accuracy.