Time Series Forecasting
Time series forecasting aims to predict future values based on historical data, crucial for diverse applications from finance to healthcare. Current research emphasizes improving model accuracy and efficiency, focusing on transformer-based architectures, state-space models like Mamba, and hybrid approaches combining their strengths, as well as exploring data augmentation and explainable AI techniques. These advancements are driving improvements in forecasting accuracy and interpretability, leading to better decision-making across various sectors and contributing to a deeper understanding of complex temporal dynamics.
Papers
Review of automated time series forecasting pipelines
Stefan Meisenbacher, Marian Turowski, Kaleb Phipps, Martin Rätz, Dirk Müller, Veit Hagenmeyer, Ralf Mikut
CoST: Contrastive Learning of Disentangled Seasonal-Trend Representations for Time Series Forecasting
Gerald Woo, Chenghao Liu, Doyen Sahoo, Akshat Kumar, Steven Hoi
ETSformer: Exponential Smoothing Transformers for Time-series Forecasting
Gerald Woo, Chenghao Liu, Doyen Sahoo, Akshat Kumar, Steven Hoi
Joint Demand Prediction for Multimodal Systems: A Multi-task Multi-relational Spatiotemporal Graph Neural Network Approach
Yuebing Liang, Guan Huang, Zhan Zhao
Optimal Latent Space Forecasting for Large Collections of Short Time Series Using Temporal Matrix Factorization
Himanshi Charotia, Abhishek Garg, Gaurav Dhama, Naman Maheshwari