Multiscale Backcast Convolution Neural Network for Traffic Flow Prediction in The Frequency Domain
With the construction of intelligent transportation systems in recent years, intelligent methods for the prediction of traffic flow are becoming more and more important, and accurate prediction plays a key role in enabling downstream scheduling algorithms. However, the accuracy of most current forec...
| Main Authors: | , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
MDPI AG
2022-11-01
|
| Series: | Applied Sciences |
| Subjects: | |
| Online Access: | https://www.mdpi.com/2076-3417/12/23/11912 |
| _version_ | 1797463668069236736 |
|---|---|
| author | Shuying Wang Yinong Zhang En Fu Shaohu Tang |
| author_facet | Shuying Wang Yinong Zhang En Fu Shaohu Tang |
| author_sort | Shuying Wang |
| collection | DOAJ |
| description | With the construction of intelligent transportation systems in recent years, intelligent methods for the prediction of traffic flow are becoming more and more important, and accurate prediction plays a key role in enabling downstream scheduling algorithms. However, the accuracy of most current forecasting algorithms remains unsatisfactory. Because traffic depends on the time of the day and varies throughout the week, such as during peak commuting periods as opposed to other times, traffic flow data show evident cyclical patterns. We capitalize on this notion and propose a multiscale convolutional feedback network for frequency prediction based on frequency angle. We combine multiscale convolution (MSC) with dilated convolution, and increase the convolutional receptive field by expanding cavity size while retaining similar parameterization costs, and achieve multiscale convolution with kernels referring to different receptive fields. At the same time, we incorporate an autoencoding module by assigning the same set of hidden features to input reconstruction and output prediction, which results in enhanced stability of features within the hidden layers. When we tested our approach on the Traffic dataset, our model achieved the best performance as assessed via the three indicators measured using mean squared error (MSE), mean absolute error (MAE), and correlation coefficient (CORR), with improvements of 3.818%, 2.472% and, 0.1515%, respectively. |
| first_indexed | 2024-03-09T17:54:01Z |
| format | Article |
| id | doaj.art-3341f679016f409987d8cf43c7741ced |
| institution | Directory Open Access Journal |
| issn | 2076-3417 |
| language | English |
| last_indexed | 2024-03-09T17:54:01Z |
| publishDate | 2022-11-01 |
| publisher | MDPI AG |
| record_format | Article |
| series | Applied Sciences |
| spelling | doaj.art-3341f679016f409987d8cf43c7741ced2023-11-24T10:27:53ZengMDPI AGApplied Sciences2076-34172022-11-0112231191210.3390/app122311912Multiscale Backcast Convolution Neural Network for Traffic Flow Prediction in The Frequency DomainShuying Wang0Yinong Zhang1En Fu2Shaohu Tang3College of Urban Rail Transit and Logistics, Beijing Union University, Beijing 100101, ChinaCollege of Urban Rail Transit and Logistics, Beijing Union University, Beijing 100101, ChinaBeijing Key Laboratory of Information Service Engineering, Beijing Union University, Beijing 100101, ChinaCollege of Urban Rail Transit and Logistics, Beijing Union University, Beijing 100101, ChinaWith the construction of intelligent transportation systems in recent years, intelligent methods for the prediction of traffic flow are becoming more and more important, and accurate prediction plays a key role in enabling downstream scheduling algorithms. However, the accuracy of most current forecasting algorithms remains unsatisfactory. Because traffic depends on the time of the day and varies throughout the week, such as during peak commuting periods as opposed to other times, traffic flow data show evident cyclical patterns. We capitalize on this notion and propose a multiscale convolutional feedback network for frequency prediction based on frequency angle. We combine multiscale convolution (MSC) with dilated convolution, and increase the convolutional receptive field by expanding cavity size while retaining similar parameterization costs, and achieve multiscale convolution with kernels referring to different receptive fields. At the same time, we incorporate an autoencoding module by assigning the same set of hidden features to input reconstruction and output prediction, which results in enhanced stability of features within the hidden layers. When we tested our approach on the Traffic dataset, our model achieved the best performance as assessed via the three indicators measured using mean squared error (MSE), mean absolute error (MAE), and correlation coefficient (CORR), with improvements of 3.818%, 2.472% and, 0.1515%, respectively.https://www.mdpi.com/2076-3417/12/23/11912traffic flow predictiontime seriesconvolutional neural networksauto-encoderintelligent transportation systems |
| spellingShingle | Shuying Wang Yinong Zhang En Fu Shaohu Tang Multiscale Backcast Convolution Neural Network for Traffic Flow Prediction in The Frequency Domain Applied Sciences traffic flow prediction time series convolutional neural networks auto-encoder intelligent transportation systems |
| title | Multiscale Backcast Convolution Neural Network for Traffic Flow Prediction in The Frequency Domain |
| title_full | Multiscale Backcast Convolution Neural Network for Traffic Flow Prediction in The Frequency Domain |
| title_fullStr | Multiscale Backcast Convolution Neural Network for Traffic Flow Prediction in The Frequency Domain |
| title_full_unstemmed | Multiscale Backcast Convolution Neural Network for Traffic Flow Prediction in The Frequency Domain |
| title_short | Multiscale Backcast Convolution Neural Network for Traffic Flow Prediction in The Frequency Domain |
| title_sort | multiscale backcast convolution neural network for traffic flow prediction in the frequency domain |
| topic | traffic flow prediction time series convolutional neural networks auto-encoder intelligent transportation systems |
| url | https://www.mdpi.com/2076-3417/12/23/11912 |
| work_keys_str_mv | AT shuyingwang multiscalebackcastconvolutionneuralnetworkfortrafficflowpredictioninthefrequencydomain AT yinongzhang multiscalebackcastconvolutionneuralnetworkfortrafficflowpredictioninthefrequencydomain AT enfu multiscalebackcastconvolutionneuralnetworkfortrafficflowpredictioninthefrequencydomain AT shaohutang multiscalebackcastconvolutionneuralnetworkfortrafficflowpredictioninthefrequencydomain |