SCINet: Time Series Modeling and Forecasting with Sample Convolution and Interaction

Abstract

One unique property of time series is that the temporal relations are largely preserved after downsampling into two sub-sequences. By taking advantage of this property, we propose a novel neural network architecture that conducts sample convolution and interaction for temporal modeling and forecasting, named SCINet. Specifically, SCINet is a recursive downsample-convolve-interact architecture. In each layer, we use multiple convolutional filters to extract distinct yet valuable temporal features from the downsampled sub-sequences or features. By combining these rich features aggregated from multiple resolutions, SCINet effectively models time series with complex temporal dynamics. Experimental results show that SCINet achieves significant forecasting accuracy improvements over both existing convolutional models and Transformer-based solutions across various real-world time series forecasting datasets. Our codes and data are available at https://github.com/cure-lab/SCINet.

arXiv commentThis paper presents a novel convolutional neural network for time series forecasting, achieving significant accuracy improvements

References in corpus (14)

• Neural Machine Translation by Jointly Learning to Align and Translate
• An Empirical Evaluation of Generic Convolutional and Recurrent Networks for Sequence Modeling
• WaveNet: A Generative Model for Raw Audio
• Time Series Forecasting With Deep Learning: A Survey
• On the Properties of Neural Machine Translation: Encoder-Decoder Approaches
• Enhancing the Locality and Breaking the Memory Bottleneck of Transformer on Time Series Forecasting
• Adaptive Graph Convolutional Recurrent Network for Traffic Forecasting
• Multi-Scale Convolutional Neural Networks for Time Series Classification
• Conditional Time Series Forecasting with Convolutional Neural Networks
• Reformer: The Efficient Transformer
• Trellis Networks for Sequence Modeling
• Machine Learning for Spatiotemporal Sequence Forecasting: A Survey
• Temporal-Clustering Invariance in Irregular Healthcare Time Series
• Temporal Latent Auto-Encoder: A Method for Probabilistic Multivariate Time Series Forecasting

Cited by in corpus (5)

• Spatial-Temporal Interactive Dynamic Graph Convolution Network for Traffic Forecasting
• Time-Series Forecasting in Smart Manufacturing Systems: An Experimental Evaluation of the State-of-the-art Algorithms
• MoGERNN: An Inductive Traffic Predictor for Unobserved Locations
• Full-Body Motion Reconstruction with Sparse Sensing from Graph Perspective
• Limited Resource Allocation in a Non-Markovian World: The Case of Maternal and Child Healthcare