Multivariate machine learning-based prediction models of freeway traffic flow under non-recurrent events
Fahad Aljuaydi, Benchawan Wiwatanapataphee, Yonghong Wu
Abstract
This paper concerns multivariate machine learning-based prediction models of freeway traffic flow under non-recurrent events. Five model architectures based on the multi-layer perceptron (MLP), convolutional neural network (CNN), long short-term memory (LSTM), CNN-LSTM and Autoencoder LSTM networks have been developed to predict traffic flow under a road crash and the rain. Using an input dataset with five features (the flow rate, the speed, and the density, road incident and rainfall) and two standard metrics (the Root Mean Square error and the Mean Absolute error), models’ performance is evaluated.
Topics & Concepts
AutoencoderMultivariate statisticsComputer scienceMean squared errorConvolutional neural networkTraffic flow (computer networking)Multilayer perceptronArtificial intelligencePerceptronArtificial neural networkRecurrent neural networkMachine learningStatisticsMathematicsComputer securityTraffic Prediction and Management TechniquesTraffic control and managementTraffic and Road Safety