Zero-shot Multivariate Time Series Forecasting Using Tabular Prior Fitted Networks

Bethmage Mayuka Jayawardhana, Nihal Sharma, Kazem Meidani, C. Bayan Bruss, Tom Goldstein, Doron L Bergman

Presentation Attendance: Yes, we will present in-person

Keywords: Time Series Forecasting, Multivariate, Prior Fitted Networks

TL;DR: We propose a framework to apply tabular foundation models (like TabPFN) directly for multivariate time series forecasting, by recasting the task as a scaler regression problem.

Abstract: Tabular foundation models, particularly Prior-data Fitted Networks like TabPFN have emerged as the leading contender in a myriad of tasks ranging from data imputation to label prediction on the tabular data format surpassing the historical successes of tree-based models. This has led to investigations on their applicability to forecasting time series data which can be formulated as a tabular problem. While recent work to this end has displayed positive results, most works have limited their treatment of multivariate time series problems to several independent univariate time series forecasting subproblems, thus ignoring any inter-channel interactions. Overcoming this limitation, we introduce a generally applicable framework for multivariate time series forecasting using tabular foundation models. We achieve this by recasting the multivariate time series forecasting problem as a *series* of scalar regression problems which can then be solved *zero-shot* by any tabular foundation model with regression capabilities. We present results of our method using the TabPFN-TS backbone and compare performance with the current state of the art tabular methods.

Track: Research Track (max 4 pages)

Submission Number: 33