Beyond Sentiment: Structured Information Extraction from Financial News

2026-07-30

AI Summary

This study addresses the limitations of conventional financial news sentiment analysis, which collapses multidimensional semantics into a single polarity score, thereby neglecting critical dimensions such as event type, scope of impact, and temporal span, leading to significant predictive information loss. To overcome this, the authors propose a structured information extraction framework that leverages LLaMA-3.1-70B to identify six non-sentiment semantic dimensions from news texts and integrates them with FinBERT-derived sentiment features. The work provides the first systematic validation of the independent predictive utility of these non-sentiment dimensions (ΔF1 = +0.019), revealing a decoupling phenomenon between sentiment and semantic content. The fused model achieves an F1 score of 0.600—significantly outperforming baseline methods—with each dimension contributing equitably to performance (14%–21%).

Abstract

Financial sentiment analysis has become a standard component in news-driven stock prediction, yet it reduces rich, multi-dimensional news articles to a single polarity score. We hypothesize that financial news encodes multiple orthogonal information dimensions---event type, impact scope, temporal horizon, and semantic confidence---that sentiment alone cannot capture, and that these dimensions carry independent predictive value. To test this hypothesis, we propose a structured information extraction framework that leverages LLaMA-3.1-70B to extract six semantic dimensions from financial news. Through large-scale experiments on 41,618 news--stock pairs from the FNSPID dataset, we find that (i) FinBERT sentiment features exhibit strong predictive power under nonlinear models (F1=0.576) but substantially weaker performance under linear models (F1=0.230), revealing a highly nonlinear sentiment--return relationship; (ii) LLM-extracted structured features, while individually weaker, capture information orthogonal to sentiment, as evidenced by a 53.5% systematic disagreement rate between the two approaches; and (iii) combining both signal sources yields F1=0.600, significantly outperforming either alone ($p < 0.0001$), with consistent improvements across all seven event types. Ablation experiments confirm that non-sentiment structural dimensions (event type, impact subject, time horizon, confidence) independently contribute $Δ\text{F1} = +0.019$ beyond FinBERT alone. Feature importance analysis reveals balanced contributions from all six extracted dimensions (14--21%), demonstrating that compressing news into a single sentiment score incurs substantial information loss. Our results suggest that the sentiment--semantics decoupling in financial text is systematic and exploitable, opening a new direction for multi-dimensional financial NLP.

Problem

• financial sentiment analysis
• structured information extraction
• multi-dimensional financial NLP
• information loss
• orthogonal information dimensions

Innovation

• structured information extraction
• financial NLP
• LLM-based feature extraction
• sentiment-semantics decoupling
• multi-dimensional news representation