Temporal Graph Learning for Pair Trading

Summary

This document describes a deep-learning framework for identifying time-varying relationships between assets for pair trading. Multi-modal Temporal Relation Graph Learning combines time-series observations with discrete features in a temporal graph. A memory-based temporal graph neural network then treats the task of finding related entities as temporal link prediction, with the goal of informing automated pair selection.

The authors report experiments on real-world datasets and say the framework performs better than alternatives, but the document does not name the datasets, benchmarks, evaluation measures, or trading costs. It also gives no details about how predicted links become entry and exit rules, how the resulting portfolio is managed, or whether performance persists out of sample. The description therefore outlines a modeling approach and its claimed empirical promise, rather than enough evidence to judge a deployable strategy.

Key ideas

• The framework targets changing relationships between assets used in pair trading.
• It combines time-series data and discrete features in a temporal graph.
• A memory-based graph neural network predicts links between entities over time.
• The authors report favorable results on real-world datasets, without providing benchmark details in this description.
• The document does not explain how link predictions translate into trading and risk rules.

Tags

# MTRGL:Effective Temporal Correlation Discerning through Multi-modal Temporal Relational Graph Learning


# MTRGL:Effective Temporal Correlation Discerning through Multi-modal Temporal Relational Graph Learning









In this study, we explore the synergy of deep learning and financial market applications, focusing on pair trading. This market-neutral strategy is integral to quantitative finance and is apt for advanced deep-learning techniques. A pivotal challenge in pair trading is discerning temporal correlations among entities, necessitating the integration of diverse data modalities. Addressing this, we introduce a novel framework, Multi-modal Temporal Relation Graph Learning (MTRGL). MTRGL combines time series data and discrete features into a temporal graph and employs a memory-based temporal graph neural network. This approach reframes temporal correlation identification as a temporal graph link prediction task, which has shown empirical success. Our experiments on real-world datasets confirm the superior performance of MTRGL, emphasizing its promise in refining automated pair trading strategies.

Shown in full with attribution under the source's licence. Licence: abstract CC0

This summary was written by Stratmill's research agent from the original; it is not a copy of the source.