Internet traffic matrix prediction with convolutional LSTM neural network

Abstract

With the rapid growing trend of Internet, prediction-based network operation optimization and management has drawn the attention from both the academia and the industry. For predicting Internet traffic, deep learning models have been proven more effective than linear and statistical models. However, some of the previous studies model the Internet traffic prediction as a multivariate time series prediction problem simply, without using the Internet traffic matrix structure. In this letter, the Internet traffic matrix prediction problem is firstly modeled as a video prediction task. Then a ConvLSTM-based Seq2Seq model named ConvLSTM-TM is proposed for predicting the traffic matrix in the next time slot. Based on three real-world traffic matrix datasets, ConvLSTM-TM outperforms five deep learning baselines with a lower prediction error. For replication, the dataset, code, and results are publicly available in a Github repository.

Traffic prediction with deep learning for vehicular communication network

“…Two advanced deep learning models are evaluated in this study, namely, LSTM and GRU, both of which are improved variants of RNNs. These models have been successful for many similar problems [13][14]. As for comparison, the normal feed-forward artificial neural network (ANN) is also used in this study.…”

Vehicle destination prediction with spatial clustering and machine learning

“…The destination coordinates in the training subset are used as the input for the DBSCAN clustering and the test subset is used for prediction performance evaluation. We adjust the threshold parameter of the minimum number of data samples required to be considered as a cluster in the DBSCAN clustering algorithm with values from (3,5,10,20), to obtain different cluster numbers for comparison. There are approximately 5800 vehicles in the dataset.…”

“…The ability to precisely predict vehicle destinations is a prerequisite for providing more applications in the Internet of Vehicles, 2,3 including content caching, 4,5 data dissemination, 6 routing 7 and scheduling. 8 While traffic prediction has been widely considered in the Internet of Vehicles with machine learning and deep learning techniques, 9,10 the existing methods still have their limitations, especially in the limited data scenario.…”

Deep learning based short‐term load forecasting incorporating calendar and weather information

“…4 In the literature, they have been proven effective for a wide collection of time series forecasting problems. 5,6,7,8 These deep learning models have also been successfully applied for short-term load forecasting with extraordinary performance. 9 For load forecasting, LSTM has been proven effective in adapting and learning the underlying complex features over time 10 and achieving the best forecasting performance.…”