Zero Shot Time Series Forecasting Using Kolmogorov Arnold Networks

Abhiroop Bhattacharya, Nandinee Haq · 2024

Read Zero Shot Time Series Forecasting Using Kolmogorov Arnold Networks by Abhiroop Bhattacharya, Nandinee Haq (2024) 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-2412-17853.txt.

Description

Bhattacharya and Haq present a forecasting model, published in 2024, that aims to predict energy prices in markets it has never seen. Day-ahead prices steer real money decisions, so accuracy matters. Models usually learn one market and stay trapped in it. The authors train their model to extract patterns that hold across markets instead. At its core sits a doubly layered forecasting network whose building blocks are a newer kind of network layer.

That layer type rests on a classic theorem: complex functions of many variables can be rebuilt from simple one-variable pieces. The design promises flexible fits with clear structure. Tested across markets, the model transfers without retraining on the new market's data. Transfer forecasting could help smaller markets with thin data histories. It also probes whether price dynamics share a common shape worldwide.

Catalog record

CREATORAbhiroop Bhattacharya, Nandinee Haq
YEAR2024
COLLECTIONPAPERS
FORMATTXT
FILE SIZE22 KB
VERSION1.0
LICENSEhttp://creativecommons.org/licenses/by/4.0/
RIGHTSCC BY
SOURCEarXiv

Source: https://arxiv.org/abs/2412-17853

Browse the full library · About the Tower of Babel