Partial multivariate transformer as a tool for cryptocurrencies time series prediction
Andrzej Tokajuk, Jarosław A. Chudziak
TL;DR
The paper tackles cryptocurrency price forecasting under extreme volatility by testing a Partial Multivariate Transformer (PMformer) that uses strategically selected feature subsets. It demonstrates that partial multivariate approaches can achieve superior statistical accuracy compared to univariate and full multivariate models. However, it also reveals a notable disconnect between lower prediction error and trading profitability, which varies by asset (e.g., BTCUSDT vs ETHUSDT). The findings call for evaluation criteria that align more closely with real-world trading objectives and for further exploration of adaptive feature selection in financial time series.
Abstract
Forecasting cryptocurrency prices is hindered by extreme volatility and a methodological dilemma between information-scarce univariate models and noise-prone full-multivariate models. This paper investigates a partial-multivariate approach to balance this trade-off, hypothesizing that a strategic subset of features offers superior predictive power. We apply the Partial-Multivariate Transformer (PMformer) to forecast daily returns for BTCUSDT and ETHUSDT, benchmarking it against eleven classical and deep learning models. Our empirical results yield two primary contributions. First, we demonstrate that the partial-multivariate strategy achieves significant statistical accuracy, effectively balancing informative signals with noise. Second, we experiment and discuss an observable disconnect between this statistical performance and practical trading utility; lower prediction error did not consistently translate to higher financial returns in simulations. This finding challenges the reliance on traditional error metrics and highlights the need to develop evaluation criteria more aligned with real-world financial objectives.