Track papers across labs, topics, and trends.
Search papers, labs, and topics across Lattice.
This study introduces a structured information extraction framework that utilizes LLaMA-3.1-70B to capture six distinct semantic dimensions from financial news articles, arguing that these dimensions provide independent predictive value beyond traditional sentiment analysis. Through analysis of 41,618 news-stock pairs, the authors reveal that while sentiment features from FinBERT show strong predictive power under nonlinear models, they perform poorly under linear models, highlighting a complex sentiment-return relationship. The combination of LLM-extracted features and sentiment analysis significantly enhances predictive performance, achieving an F1 score of 0.600, thus demonstrating that reducing financial news to a single polarity score leads to substantial information loss.
Extracting six semantic dimensions from financial news reveals that traditional sentiment analysis misses critical predictive signals, leading to a 53.5% disagreement rate with sentiment-based models.
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 $\Delta\text{F1} = +0.019$ 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.