TRACE: Temporal Rule-Anchored Chain-of-Evidence on Knowledge Graphs for Interpretable Stock Movement Prediction
arXiv:2603.12500v1 Announce Type: cross Abstract: We present a Temporal Rule-Anchored Chain-of-Evidence (TRACE) on knowledge graphs for interpretable stock movement prediction that unifies symbolic relational priors, dynamic graph exploration, and LLM-guided decision making in a single end-to-end pipeline. The approach performs rule-guided multi-hop exploration restricted to admissible relation sequences, grounds candidate reasoning chains in contemporaneous news, and aggregates fully grounded evidence into auditable \texttt{UP}/\texttt{DOWN} verdicts with human-readable paths connecting text and structure. On an S\&P~500 benchmark, the method achieves 55.1\% accuracy, 55.7\% precision, 71.5\% recall, and 60.8\% F1, surpassing strong baselines and improving recall and F1 over the best graph baseline under identical evaluation. The gains stem from (i) rule-guided exploration that focuses search on economically meaningful motifs rather than arbitrary walks, and (ii) text-grounded consol
arXiv:2603.12500v1 Announce Type: cross Abstract: We present a Temporal Rule-Anchored Chain-of-Evidence (TRACE) on knowledge graphs for interpretable stock movement prediction that unifies symbolic relational priors, dynamic graph exploration, and LLM-guided decision making in a single end-to-end pipeline. The approach performs rule-guided multi-hop exploration restricted to admissible relation sequences, grounds candidate reasoning chains in contemporaneous news, and aggregates fully grounded evidence into auditable \texttt{UP}/\texttt{DOWN} verdicts with human-readable paths connecting text and structure. On an S\&P~500 benchmark, the method achieves 55.1\% accuracy, 55.7\% precision, 71.5\% recall, and 60.8\% F1, surpassing strong baselines and improving recall and F1 over the best graph baseline under identical evaluation. The gains stem from (i) rule-guided exploration that focuses search on economically meaningful motifs rather than arbitrary walks, and (ii) text-grounded consolidation that selectively aggregates high-confidence, fully grounded hypotheses instead of uniformly pooling weak signals. Together, these choices yield higher sensitivity without sacrificing selectivity, delivering predictive lift with faithful, auditably interpretable explanations.
Executive Summary
The article introduces TRACE, a novel approach for interpretable stock movement prediction using knowledge graphs. TRACE combines symbolic relational priors, dynamic graph exploration, and LLM-guided decision making to achieve 55.1% accuracy, 55.7% precision, 71.5% recall, and 60.8% F1 on an S&P 500 benchmark. The method's gains stem from rule-guided exploration and text-grounded consolidation, yielding higher sensitivity without sacrificing selectivity. TRACE delivers predictive lift with faithful, auditably interpretable explanations, surpassing strong baselines and improving recall and F1 over the best graph baseline.
Key Points
- ▸ Introduction of TRACE, a Temporal Rule-Anchored Chain-of-Evidence approach
- ▸ Combination of symbolic relational priors, dynamic graph exploration, and LLM-guided decision making
- ▸ Achievement of 55.1% accuracy, 55.7% precision, 71.5% recall, and 60.8% F1 on an S&P 500 benchmark
Merits
Improved Predictive Accuracy
TRACE achieves higher accuracy and recall compared to strong baselines and the best graph baseline.
Interpretable Explanations
TRACE provides faithful, auditably interpretable explanations for stock movement predictions.
Demerits
Complexity
The TRACE approach may be complex to implement and require significant computational resources.
Data Quality
The performance of TRACE may be sensitive to the quality of the input data and the knowledge graph.
Expert Commentary
The introduction of TRACE marks a significant advancement in the development of interpretable stock movement prediction models. By combining symbolic relational priors, dynamic graph exploration, and LLM-guided decision making, TRACE achieves impressive predictive accuracy and provides faithful explanations for its predictions. However, the complexity of the approach and the sensitivity to data quality may limit its widespread adoption. Further research is needed to address these challenges and explore the potential applications of TRACE in financial markets.
Recommendations
- ✓ Further evaluation of TRACE on diverse datasets and financial markets
- ✓ Investigation of techniques to simplify the implementation and reduce the computational resources required by TRACE
