CryptoGAT: Are Time Series Models Effective for Cryptocurrency Forecasting?

2026-06-25

AI Summary

Traditional time series models struggle to effectively capture predictable signals in cryptocurrency prices under extreme volatility. This work reframes cryptocurrency price forecasting as a cross-asset graph modeling problem rather than a conventional sequential modeling task and introduces CryptoGAT, a lightweight graph attention network, for prediction. The study is the first to reveal fundamental differences between stocks and cryptocurrencies in terms of signal predictability and cross-asset dependencies, thereby establishing a novel graph-based paradigm for cryptocurrency forecasting. Experimental results on real-world datasets demonstrate that the proposed approach significantly outperforms mainstream time series models—including LSTM, GRU, and Transformer—validating the superiority of the graph-based modeling framework.

Abstract

Cryptocurrency price prediction is a significant challenge in quantitative investment. In recent years, time series models have made significant progress in financial forecasting tasks, especially in the stock market. Despite the growing performance over the past few years, we question the validity of this line of research in cryptocurrency prediction. Specifically, time series models (e.g., LSTM, GRU, and Transformers) are effective at extracting temporal relationships in stock market data. However, in pure price-based cryptocurrency prediction, facing data with extreme volatility and wild swings, time series models have difficulty learning effective information. To validate our claim, we propose CryptoGAT, a lightweight Graph Attention Network that recasts cryptocurrency pure price prediction as a cross-asset graph problem rather than a temporal modeling task. Extensive experiments on real cryptocurrency benchmarks demonstrate that our proposed CryptoGAT outperforms various state-of-the-art forecasting methods with a notable margin. Moreover, we conduct comprehensive empirical studies to explore the fundamental differences exposed by time series models in stock and cryptocurrency prediction: differences in predictability of the signal and cross-asset dependencies. This finding opens up new research directions for the cryptocurrency pure price prediction task and inspires further graph-based exploration in the field. The source code is available at https://github.com/FanBroWell/CryptoGAT

Problem

• cryptocurrency forecasting
• time series models
• price prediction
• volatility
• cross-asset dependencies

Innovation

• Graph Attention Network
• Cryptocurrency Forecasting
• Cross-asset Dependency
• Time Series Limitations
• Pure Price Prediction