Adaptive Temporal Fusion Transformers for Cryptocurrency Price Prediction

2025-09-06

AI Summary

To address the low accuracy of short-term cryptocurrency price forecasting caused by high volatility and non-stationarity, this paper proposes an Adaptive Temporal Fusion Transformer (ATFT) framework. The method introduces three key innovations: (1) a dynamic subsequence partitioning mechanism based on relative extrema; (2) subsequence categorization according to fixed-pattern endings to decouple and model distinct market response patterns; and (3) a multi-branch Temporal Fusion Transformer architecture that integrates initial trend features to enable conditional forecasting. Evaluated on 10-minute ETH-USDT data, ATFT achieves statistically significant improvements over standard TFT and LSTM in both prediction error (MAE/RMSE) and simulated trading profitability. These results empirically validate the effectiveness of pattern-aware modeling for high-frequency cryptocurrency asset forecasting.

Technology Category

• Data Mining & Knowledge Management: Mining of Spatial, Temporal or Spatio-Temporal Data
• Machine Learning: Time-Series/Data Streams
• Planning, Routing, and Scheduling: Temporal Planning

Application Category

• Graph Algorithms and Modeling for the Web: Efficient manipulation of static and dynamic Web-related graphs
• Economics, Online Markets and Human Computation: Economic aspects of blockchain and cryptocurrencies
• Security and Privacy: Cryptocurrency and smart contracts

Abstract

Precise short-term price prediction in the highly volatile cryptocurrency market is critical for informed trading strategies. Although Temporal Fusion Transformers (TFTs) have shown potential, their direct use often struggles in the face of the market's non-stationary nature and extreme volatility. This paper introduces an adaptive TFT modeling approach leveraging dynamic subseries lengths and pattern-based categorization to enhance short-term forecasting. We propose a novel segmentation method where subseries end at relative maxima, identified when the price increase from the preceding minimum surpasses a threshold, thus capturing significant upward movements, which act as key markers for the end of a growth phase, while potentially filtering the noise. Crucially, the fixed-length pattern ending each subseries determines the category assigned to the subsequent variable-length subseries, grouping typical market responses that follow similar preceding conditions. A distinct TFT model trained for each category is specialized in predicting the evolution of these subsequent subseries based on their initial steps after the preceding peak. Experimental results on ETH-USDT 10-minute data over a two-month test period demonstrate that our adaptive approach significantly outperforms baseline fixed-length TFT and LSTM models in prediction accuracy and simulated trading profitability. Our combination of adaptive segmentation and pattern-conditioned forecasting enables more robust and responsive cryptocurrency price prediction.

Problem

• Adaptive TFTs for volatile cryptocurrency short-term price prediction
• Dynamic subseries segmentation using pattern-based categorization
• Specialized models for different market phases to improve accuracy

Innovation

• Adaptive TFT with dynamic subseries segmentation
• Pattern-based categorization for specialized model training
• Relative maxima thresholding to filter noise