An adaptive network with consecutive and intertwined slices for real-world time-series forecasting

Li Shen, Yuning Wei, Yangzhu Wang, Hongguang Li

Abstract: Highlights•A novel forecasting network achieving state-of-the-art while free of hyper-parameter fine-tuning.•The first forecasting network that can adaptively determine input sequence lengths.•A novel feature selection strategy for interpretable variable treatment and input sequence segmentation.•Better performances than seven baselines on eleven benchmarks with 22.8% MSE reduction in general.