vLinear: A Powerful Linear Model for Multivariate Time Series Forecasting

Wenzhen Yue, Ruohao Guo, Ji Shi, Zihan Hao, Shiyu Hu, Xianghua Ying · 2026

Read vLinear: A Powerful Linear Model for Multivariate Time Series Forecasting by Wenzhen Yue, Ruohao Guo, Ji Shi, Zihan Hao, Shiyu Hu, Xianghua Ying (2026) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2601-13768.txt.

Introduction

Wenzhen Yue and co-authors present a lean linear model for forecasting many time series at once in 2026. Their model, vLinear, targets multivariate series, where many values move together. Top forecasters often use self-attention to catch links between series. That step can cost in line with the square of the series count.

vLinear has two parts: a vecTrans module and a WFMLoss objective. The design aims for strong forecasts with linear, not quadratic, cost. Many series arise in energy, traffic, and finance work. A fast, true forecaster for them has broad daily use.

Catalog record

CREATORWenzhen Yue, Ruohao Guo, Ji Shi, Zihan Hao, Shiyu Hu, Xianghua Ying
YEAR2026
COLLECTIONPAPERS
FORMATTXT
FILE SIZE146 KB
VERSION1.0
LICENSEhttp://creativecommons.org/licenses/by/4.0/
RIGHTSCC BY
SOURCEarXiv

Source: https://arxiv.org/abs/2601-13768

Browse the full library · About the Tower of Babel