PROJECT TITLE :

Traffic Flow Prediction With Big Data: A Deep Learning Approach

ABSTRACT:

Correct and timely traffic flow info is very important for the successful deployment of intelligent transportation systems. Over the previous couple of years, traffic information are exploding, and we have really entered the age of huge information for transportation. Existing traffic flow prediction methods mainly use shallow traffic prediction models and are still unsatisfying for several real-world applications. This scenario conjures up us to rethink the traffic flow prediction drawback based mostly on deep architecture models with big traffic information. In this paper, a completely unique deep-learning-based mostly traffic flow prediction technique is proposed, which considers the spatial and temporal correlations inherently. A stacked autoencoder model is used to be told generic traffic flow features, and it is trained in a very greedy layerwise fashion. To the simplest of our knowledge, this can be the first time that a deep design model is applied using autoencoders as building blocks to represent traffic flow features for prediction. Moreover, experiments demonstrate that the proposed methodology for traffic flow prediction has superior performance.