ANALYSIS AND COMPARISON OF LONG SHORT-TERM MEMORY NETWORKS SHORT-TERM TRAFFIC PREDICTION PERFORMANCE
Long short-term memory networks (LSTM) produces promising results in the prediction of traffic flows. However, LSTM needs large numbers of data to produce satisfactory results. Therefore, the effect of LSTM training set size on performance and optimum training set size for short-term traffic flow pr...
| Main Author: | |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
Silesian University of Technology
2020-08-01
|
| Series: | Scientific Journal of Silesian University of Technology. Series Transport |
| Subjects: | |
| Online Access: | http://sjsutst.polsl.pl/archives/2020/vol107/019_SJSUTST107_2020_Dogan.pdf |
| _version_ | 1818365535640879104 |
|---|---|
| author | Erdem DOGAN |
| author_facet | Erdem DOGAN |
| author_sort | Erdem DOGAN |
| collection | DOAJ |
| description | Long short-term memory networks (LSTM) produces promising results in the prediction of traffic flows. However, LSTM needs large numbers of data to produce satisfactory results. Therefore, the effect of LSTM training set size on performance and optimum training set size for short-term traffic flow prediction problems were investigated in this study. To achieve this, the numbers of data in the training set was set between 480 and 2800, and the prediction performance of the LSTMs trained using these adjusted training sets was measured. In addition, LSTM prediction results were compared with nonlinear autoregressive neural networks (NAR) trained using the same training sets. Consequently, it was seen that the increase in LSTM's training cluster size increased performance to a certain point. However, after this point, the performance decreased. Three main results emerged in this study: First, the optimum training set size for LSTM significantly improves the prediction performance of the model. Second, LSTM makes short-term traffic forecasting better than NAR. Third, LSTM predictions fluctuate less than the NAR model following instant traffic flow changes. |
| first_indexed | 2024-12-13T22:21:49Z |
| format | Article |
| id | doaj.art-a476dd35a225456ebc140ad63803e90e |
| institution | Directory Open Access Journal |
| issn | 0209-3324 2450-1549 |
| language | English |
| last_indexed | 2024-12-13T22:21:49Z |
| publishDate | 2020-08-01 |
| publisher | Silesian University of Technology |
| record_format | Article |
| series | Scientific Journal of Silesian University of Technology. Series Transport |
| spelling | doaj.art-a476dd35a225456ebc140ad63803e90e2022-12-21T23:29:21ZengSilesian University of TechnologyScientific Journal of Silesian University of Technology. Series Transport0209-33242450-15492020-08-011071071932https://doi.org/10.20858/sjsutst.2020.107.2ANALYSIS AND COMPARISON OF LONG SHORT-TERM MEMORY NETWORKS SHORT-TERM TRAFFIC PREDICTION PERFORMANCEErdem DOGANLong short-term memory networks (LSTM) produces promising results in the prediction of traffic flows. However, LSTM needs large numbers of data to produce satisfactory results. Therefore, the effect of LSTM training set size on performance and optimum training set size for short-term traffic flow prediction problems were investigated in this study. To achieve this, the numbers of data in the training set was set between 480 and 2800, and the prediction performance of the LSTMs trained using these adjusted training sets was measured. In addition, LSTM prediction results were compared with nonlinear autoregressive neural networks (NAR) trained using the same training sets. Consequently, it was seen that the increase in LSTM's training cluster size increased performance to a certain point. However, after this point, the performance decreased. Three main results emerged in this study: First, the optimum training set size for LSTM significantly improves the prediction performance of the model. Second, LSTM makes short-term traffic forecasting better than NAR. Third, LSTM predictions fluctuate less than the NAR model following instant traffic flow changes.http://sjsutst.polsl.pl/archives/2020/vol107/019_SJSUTST107_2020_Dogan.pdfdeep learningtraffic flowshort-termpredictionlstmnonlinear autoregressivetraining set size |
| spellingShingle | Erdem DOGAN ANALYSIS AND COMPARISON OF LONG SHORT-TERM MEMORY NETWORKS SHORT-TERM TRAFFIC PREDICTION PERFORMANCE Scientific Journal of Silesian University of Technology. Series Transport deep learning traffic flow short-term prediction lstm nonlinear autoregressive training set size |
| title | ANALYSIS AND COMPARISON OF LONG SHORT-TERM MEMORY NETWORKS SHORT-TERM TRAFFIC PREDICTION PERFORMANCE |
| title_full | ANALYSIS AND COMPARISON OF LONG SHORT-TERM MEMORY NETWORKS SHORT-TERM TRAFFIC PREDICTION PERFORMANCE |
| title_fullStr | ANALYSIS AND COMPARISON OF LONG SHORT-TERM MEMORY NETWORKS SHORT-TERM TRAFFIC PREDICTION PERFORMANCE |
| title_full_unstemmed | ANALYSIS AND COMPARISON OF LONG SHORT-TERM MEMORY NETWORKS SHORT-TERM TRAFFIC PREDICTION PERFORMANCE |
| title_short | ANALYSIS AND COMPARISON OF LONG SHORT-TERM MEMORY NETWORKS SHORT-TERM TRAFFIC PREDICTION PERFORMANCE |
| title_sort | analysis and comparison of long short term memory networks short term traffic prediction performance |
| topic | deep learning traffic flow short-term prediction lstm nonlinear autoregressive training set size |
| url | http://sjsutst.polsl.pl/archives/2020/vol107/019_SJSUTST107_2020_Dogan.pdf |
| work_keys_str_mv | AT erdemdogan analysisandcomparisonoflongshorttermmemorynetworksshorttermtrafficpredictionperformance |