MEHGT-LKG: Multimodal Edge-enhanced Heterogeneous Graph Transformer with LLM-driven Knowledge Graph for Stock Trend Prediction

ICLR 2026 Conference Submission22589 Authors

Keywords: Heterogeneous graph model, Large language model, Multimodal fusion, Quantitative investment

TL;DR: We present MEHGT-LKG, a multimodal edge-enhanced heterogeneous graph transformer leveraging LLM-driven knowledge graphs to integrate financial events, market indicators, and heterogeneous relations for accurate stock trend prediction.

Abstract: Stock trend prediction plays a central role in optimal investment decision-making, and has attracted extensive research from both investors and institutions. Although recent studies have employed graph structures to model the complex relationships among financial entities, the corresponding models fail to efficiently capture semantically rich edge features across heterogeneous entities, thereby limiting the ability to fuse and align multimodal data such as market indicators, financial events, and heterogeneous graph structure. Therefore, in this paper, we propose a Multimodal Edge-Enhanced Heterogeneous Graph Transformer with LLM-driven Knowledge Graphs (MEHGT-LKG) for stock trend prediction. Specifically, we first fine-tune a large language model (LLM) by using instruction tuning datasets to design a financial event-centric knowledge extraction agent (FinEX). Subsequently, we encode the structured tuples generated from FinEX into financial event-centric knowledge graphs (FEKGs) and then construct multimodal heterogeneous graphs by incorporating multimodal information. Finally, we design a Multimodal Edge-Enhanced Heterogeneous Graph Transformer (MEHGT) to fully encode a series of semantically enriched multimodal heterogeneous graphs spanning different time horizons. MEHGT models edge-level features through type-specific encoders and integrates them into both multi-head attention and message propagation, significantly enriching the representation of relational semantics and target nodes. Extensive experimental results and trading simulations on multiple real-world datasets demonstrate the superior performance of the proposed approach beyond other state-of-the-art models.

Primary Area: learning on graphs and other geometries & topologies

Submission Number: 22589