Linear Forecasting

Linear forecasting, aiming to predict future values in time series data using linear models, remains a surprisingly competitive approach despite the rise of complex deep learning methods. Current research focuses on improving linear model performance through techniques like feature normalization and hierarchical partitioning, as well as exploring hybrid approaches that selectively incorporate deep learning models for specific, challenging predictions to enhance interpretability. This renewed interest in linear methods stems from their superior performance in certain contexts, particularly regarding interpretability and computational efficiency, impacting various fields like traffic forecasting and weather prediction.

Papers

September 16, 2024

AALF: Almost Always Linear Forecasting
Matthias Jakobs, Thomas Liebig
Deep Learning Model Time Series Forecasting Forecasting Performance Interpretable Forecast Online Model Selection Linear Forecasting

June 18, 2024

XXLTraffic: Expanding and Extremely Long Traffic Dataset for Ultra-Dynamic Forecasting Challenges
Du Yin, Hao Xue, Arian Prabowo, Shuang Ao, Flora Salim
Traffic Forecasting Traffic Datasets Linear Forecasting Traffic Forecasting Datasets

June 8, 2024

CMamba: Channel Correlation Enhanced State Space Models for Multivariate Time Series Forecasting
Chaolv Zeng, Zhanyu Liu, Guanjie Zheng, Linghe Kong
State Space Model Multivariate Time Series Mamba in Mamba Multivariate Time Series Forecasting Complex Time Series Linear Forecasting

May 22, 2024

PDMLP: Patch-based Decomposed MLP for Long-Term Time Series Forecasting
Peiwang Tang, Weitai Zhang
Transformer Architecture Long Term Time Series Forecasting Learning Decomposition Linear Forecasting

March 21, 2024

An Analysis of Linear Time Series Forecasting Models
William Toner, Luke Darlow
General Analysis Time Series Forecasting Model Generalization Linear Model Model Architecture Linear Forecasting

February 15, 2024

SAMformer: Unlocking the Potential of Transformers in Time Series Forecasting with Sharpness-Aware Minimization and Channel-Wise Attention
Romain Ilbert, Ambroise Odonnat, Vasilii Feofanov, Aladin Virmaux, Giuseppe Paolo, Themis Palpanas, Ievgen Redko
Transformer Megatron Decepticons Time Series Forecasting Channel Attention Sharpness Aware Minimization Long Term Forecasting Sharpness Aware Optimization Linear Forecasting Low Generalization

October 10, 2023

iTransformer: Inverted Transformers Are Effective for Time Series Forecasting
Yong Liu, Tengge Hu, Haoran Zhang, Haixu Wu, Shiyu Wang, Lintao Ma, Mingsheng Long
Transformer Megatron Decepticons Transformer Based Transformer Architecture Time Series Forecasting Covariate Representation Linear Forecasting

May 22, 2023

Hierarchical Partitioning Forecaster
Christopher Mattern
Precipitation Nowcasting Sequential Prediction Hierarchical Forecasting Linear Forecasting