This work examines the pipeline on Russell 2000 equities under three stock-selection regimes and suggests that stock-selection regime and allocator choice matter at least as much as the sentiment model, and that separating firm-specific and macro-exposure triggers is more informative than requiring both to fire simultaneously.
Abstract
Large language models can extract richer signals from financial news than fixed sentiment lexicons, and recent work has explored feeding such signals into portfolio construction. We study an uncertainty-aware construction that feeds model-predicted risk -- decomposed into aleatoric and epistemic components -- directly into the covariance matrix of portfolio allocators, rather than treating portfolio risk as fixed or adjusting only expected returns. We evaluate the pipeline on Russell 2000 equities under three stock-selection regimes: a pure-alpha trigger that isolates abnormal stock moves not explained by macro indicators, a pure-beta trigger that captures macro-indicator moves before the stock itself fires, and a beta trigger in which both channels agree. Across the full holding-period grid, the separated pure-alpha and pure-beta legs usually dominate the beta intersection on Sharpe and return. Two horizons are especially informative. At one day, pure beta can work under low and moderate transaction costs because it captures immediate lead-lag spillovers from liquid macro and sector indicators into exposed small-cap stocks, but this advantage disappears at 100 bps when turnover and microstructure noise dominate. At 40 days, pure beta works for a different reason: slower macro repricing overtakes the firm-specific pure-alpha channel. The strongest conservative row is pure beta with GPT-4o mini sentiment, a Student-t target, a 40-day holding period, and risk parity allocation, reaching Sharpe 2.33 at 100 bps. The results suggest that stock-selection regime and allocator choice matter at least as much as the sentiment model, and that separating firm-specific and macro-exposure triggers is more informative than requiring both to fire simultaneously.
This study examines whether measurable features of public financial news can help explain short-term stock return volatility. Public announcements, financial news, institutional opinions, social media heat and market data are converted into five variables: sentiment, topic type, popularity, publication timing and market controls. Based on an initial backtesting sample of twenty anonymized events, the paper applies event-window calculation, group comparison and a simplified regression framework to observe T+1, T+3 and T+5 market reactions. The early results show that negative news, high-heat news and core-topic news related to performance, regulation, policy or capital flows are more closely associated with stronger T+3 volatility. However, abnormal returns do not show a stable direction across groups. These findings suggest that text quantification is more useful for identifying information shocks, risk attention and review priority than for producing independent trading signals. The study also provides a basic structure for future expansion to a larger sample with clearer stock codes, information sources and release times.
This paper examines the relationship between Federal Open Market Committee (FOMC) communication surprises, global risk sentiment, and net portfolio debt inflows to twelve major emerging market economies over the period 2000–2024. Exploiting a high-frequency U.S. Monetary Policy Event-Study Database, we estimate panel fixed-effects regressions and local projections at quarterly frequency. We find that global risk sentiment, proxied by the VIX, is a robust and persistent driver of emerging market capital flows, while Fed communication surprises are statistically insignificant in normal times and in the 2022–2024 tightening cycle. A striking exception is the 2013 taper tantrum—the episode of severe capital outflow pressure triggered by Chairman Bernanke’s May 2013 congressional testimony signalling a possible tapering of asset purchases. Regime interaction tests reveal a large, highly significant negative effect of communication surprises on flows during this episode alone, with no comparable effect in 2022. Local projections confirm that the taper tantrum generated a sharp initial outflow followed by partial reversal, while VIX effects are contemporaneous but not persistent. We empirically test for market learning, finding that reduced sensitivity to Fed communication reflects a discrete recalibration after the 2013 shock rather than a gradual learning process. Regarding capital flows, the taper tantrum is clearly the exception, not the rule.
Colin Ellis· International Journal of Fin...· 0 citations
Accurately predicting stock returns remains one of the core challenges in quantitative finance, exacerbated by nonlinear market dynamics, irrational investor behavior, and the stochasticity of financial time series. This paper proposes a novel hybrid framework—Bayesian Optimized Long Short-Term Memory Network Based on Sentiment Enhancement Technical Indicators (BOLSTM-SATI), which combines a Long Short-Term Memory (LSTM) neural network with classic technical analysis signals and multi-source investor sentiment indices from the CSI 300 index constituents. This study constructs a composite feature space including MACD, RSI, Bollinger Bands, volume balance indicator, and Average True Range (ATR). Investor sentiment is quantified through a Bayesian update mechanism that integrates news text sentiment scores obtained through FinBERT and social media sentiment indices from the Eastmoney Guba forum. To mitigate overfitting and improve generalization ability, we employ a Bayesian optimization method based on a tree-structured Parzen estimator (TPE) for LSTM hyperparameter search. Robustness to perturbations is verified through Gaussian noise injection experiments. Empirical evaluation using daily data from January 2015 to December 2022 shows that BOLSTM-SATI achieves a directional accuracy of 67.3%, a mean absolute error (MAE) of 0.0082, and a Sharpe ratio of 2.41 in backtesting simulations, outperforming baseline models including ARIMA, traditional LSTM, XGBoost, and Transformer variants. The results confirm that incorporating sentiment dynamics into technical feature learning can significantly improve prediction accuracy and the profitability of trading strategies.
Bairui Peng· Advances in Economics, Manag...· 0 citations
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.
Daohan Zhu, Sitong Ge, Ruofei Wang et al.· 0 citations
Sentiment analysis (SA) of natural language text has become as a powerful instrument for enhancing financial market predictions. Quarterly reports from companies, in particular, offer a rich source of data for sentiment analysis, providing key insights into a company’s performance, strategic actions, and future prospects. These reports can significantly influence investor decisions regarding asset investments. Notwithstanding the potential, prior research has not investigated sentiment analysis concerning these resources in portfolio optimization. To fill this void, we propose an innovative three-stage approach to constructing stock portfolios. In the first stage, we perform sentiment analysis on companies’ quarterly reports using the FinBERT model to assess the sentiment surrounding each company. In the second stage, we utilize a Long-Short-Term Memory (LSTM) model for forecasting future prices, which enables the calculation of expected returns and the covariance matrix. In final stage, we present a three-objective portfolio optimization model that incorporates risk, return, and sentiment-derived trend features. We solve this model using the Weighted Goal Programming (WGP) method. Our results indicate that the proposed model effectively supports portfolio optimization. Moreover, the model is implemented using data from companies that are part of the Dow Jones Industrial Average (DJIA), and findings demonstrate high accuracy, confirming the practical potential of the proposed approach.
It is concluded that the genuinely difficult part of deployment is tail-event detection rather than directional forecasting, and that a credible evaluation must rely on event-sensitive metrics and time-aware validation; without them, performance under heavy imbalance is easily overstated.
Rishov Nag, Sankhayan Choudhury· Innovations in Systems and S...· 0 citations