Research
On-device research index

arXiv research

A locally-built, LLM-digested index of recent arXiv papers in quant finance, geometry/topology, and statistical ML — keyword search served straight from SQLite on this machine.

168,694 papers · 148 categories

Trend · papers per month

25507499 · Feb 202019922001200920172026
48 results for sentiment extraction

Study evaluates LLMs for predicting Chinese stock movements using financial news sentiments.

problem Evaluating LLMs' ability to predict stock price movements using financial news sentiments.
method Standardized experimental procedure with three LLMs, each with unique performance enhancement methods.
result Developed quantitative trading strategies and conducted back-tests to assess LLMs' performance.

The paper analyzes sports commentary to automatically recognize events and extract insights.

problem Automatically recognizing and categorizing major actions in sports events from commentary.
method Used multiple Natural Language Processing techniques for classification and sentiment analysis.
result Identified insights from analyzing live sport commentaries and classifying major actions.

LLM extracts actionable insights from customer reviews.

problem Extracting actionable insights from customer reviews.
method Large language model approach distinguishing perceptual attributes from actionable features.
result High consistency and predictive validity of LLM insights compared to human coders.

Sentiment analysis of DAX40 stocks improves performance by 5.38% annually.

problem Creating a more responsive stock market index using sentiment analysis.
method Extract sentiment from news articles, adjust index weights based on sentiment, compare performance to existing indices.
result Sentiment index outperforms DAX40 by 7.51% annually, adjusted for costs.

Anonymization reduces economic signal extraction from financial texts.

problem Reducing meaningful economic signals from financial texts due to anonymization.
method Analyzed the impact of anonymization on textual understanding and economic signal extraction.
result Information loss due to anonymization is severe and pervasive, outweighing its benefits in certain financial applications.

GPT-4 improves stock price prediction from microblogging sentiments.

problem Improving stock price prediction using sentiment analysis of microblogs.
method Developed a novel method for contextual sentiment analysis using GPT-4, fine-tuning prompts for better accuracy.
result GPT-4 outperformed BERT in predicting stock price movements, achieving a peak accuracy of 71.47%.

The study improves sentiment analysis of 10-K filings, revealing aggregation effects on accuracy and correlation with market outcomes.

problem Lack of sentiment analysis for 10-K filings, particularly for risk disclosures.
method Supervised lexicon-learning approach applied to 10-K filings and Item 1A risk-factor sections, trained against return and volatility labels at different levels of aggregation.
result Sentiment analysis of Item 1A sections performs better at the individual-firm level, while full-filing text is more accurate at sector and portfolio levels.

A new framework predicts stock movements using news sentiment and relational data.

problem Predicting stock prices from textual information is challenging due to market uncertainty and natural language complexity.
method Multi-Graph Recurrent Network (MGRN) combining textual sentiment from financial news and relational data.
result The model outperforms benchmarks in predicting stock movements.

Aspect-Based Sentiment Analysis (ABSA) deals with the extraction of sentiments and their targets. Collecting labeled data for this task in order to help neural networks generalize better can be laborious and time-consuming. As an alternative, similar data to the real-world examples can be produced artificially through …

2020-01-30abs ↗pdf ↗

To understand the relationship between news sentiment and company stock price movements, and to better understand connectivity among companies, we define an algorithm for measuring sentiment-based network risk. The algorithm ranks companies in networks of co-occurrences, and measures sentiment-based risk, by calculatin…

2017-06-19abs ↗pdf ↗

This study uses NLP to predict stock performance based on analyst reports.

problem Predicting stock performance using textual information from analyst reports.
method Natural language processing (NLP) and a customized BERT deep learning model for Chinese text.
result Strong positive sentiment in analyst reports increases excess return and intraday volatility, while strong negative sentiment increases volatility and trading volume but decreases excess return.

News sentiment in U.S. economic newspapers has become more persistent over 45 years.

problem Understanding the temporal dynamics of U.S. economic news sentiment over time.
method Daily economic news sentiment index from 1980-2025, analyzed using sentiment indexes.
result News sentiment states have become more persistent, with longer residence times in optimistic or pessimistic regimes.

Hybrid model predicts stock prices using online forum sentiments and popularity.

problem Predicting stock prices accurately considering investor sentiment.
method XLNET for sentiment analysis, BiLSTM-highway model integration, combining post popularity.
result Hybrid model outperforms traditional methods in stock price prediction.

This paper optimizes cryptocurrency portfolios by integrating sentiment analysis with technical indicators.

problem Effective portfolio management in volatile cryptocurrency markets.
method Dynamic portfolio strategy using technical indicators and sentiment analysis.
result The integrated approach outperforms traditional benchmarks and achieves stronger risk-adjusted returns.

Improved crude oil price forecasting using multi-dimensional LLM sentiment signals.

problem Challenges in predicting crude oil prices due to unstructured news.
method Extracted five sentiment dimensions from GPT-4o, Llama 3.2-3b, and FinBERT models on energy-sector news articles.
result Combining GPT-4o and FinBERT yields the best predictive performance for weekly WTI crude oil futures returns.

MANA-Net improves market predictions by dynamically weighting news sentiments.

problem Aggregated Sentiment Homogenization in financial news data.
method Dynamic market-news attention mechanism to aggregate sentiments.
result MANA-Net outperforms recent market prediction methods by 1.1% Profit & Loss and 0.252 daily Sharpe ratio.

Due to the increased availability of online reviews, sentiment analysis had been witnessed a booming interest from the researchers. Sentiment analysis is a computational treatment of sentiment used to extract and understand the opinions of authors. While many systems were built to predict the sentiment of a document or…

2019-07-26abs ↗pdf ↗

This study improves stock price forecasting by analyzing daily news sentiment.

problem Improving stock price forecasting accuracy using news sentiment.
method Data collection, preprocessing, and sentiment analysis of NITY50 stocks' news.
result LSTM models with sentiment scores outperform without them in forecasting stock prices.

Aspect based Sentiment Analysis is a major subarea of sentiment analysis. Many supervised and unsupervised approaches have been proposed in the past for detecting and analyzing the sentiment of aspect terms. In this paper, a graph-based semi-supervised learning approach for aspect term extraction is proposed. In this a…

2020-02-20abs ↗pdf ↗

This study examines investor sentiment's impact on stock market liquidity and volatility using deep learning and TVP-VAR models.

problem Investor sentiment's impact on stock market liquidity and volatility.
method Deep learning BERT model for sentiment extraction and TVP-VAR model for time-varying analysis.
result Investor sentiment has a stronger impact on stock market liquidity and volatility, with more pronounced effects in short-term shocks.

Study shows negative war news correlates with increased stock market volatility.

problem Understanding the impact of geopolitical events on financial markets.
method Used BERT model for sentiment analysis and GARCH model for volatility forecasting.
result Negative news sentiment during geopolitical crises is associated with increased stock market volatility.

Study shows publicly available news impacts financial markets.

problem Impact of publicly available news on financial markets.
method Extracted news from Common Crawl, identified relevant companies, used sentiment analysis and information theory.
result Publicly available news has significant impact on financial markets.

Improved sentiment analysis in Korean finance using masked PLMs.

problem Fine-grained sentiment analysis in non-English finance literature is lacking due to limited annotated data.
method Developed KorFinASC dataset and applied TGT-Masking to PLMs to remove non-stationary knowledge.
result Improved classification accuracy by 22.63% on KorFinASC.

Study uses ML to predict currency and bond returns from news sentiment.

problem Predicting financial returns from news sentiment.
method Pretrained FinBERT model on finance-specific language, XGBoost classifier, SHAP for interpretability.
result XGBoost strategy outperforms benchmarks with Sharpe ratios > 5.

The paper predicts Bitcoin prices using machine learning and sentiment analysis.

problem Predicting the future price of Bitcoin in USD.
method Applied supervised machine learning and sentiment analysis to Twitter and Reddit data.
result LSTM models with multi-feature analysis outperformed ARIMA models in predicting Bitcoin prices.

Study predicts market bubbles using machine learning and financial news sentiment.

problem Predicting market bubbles in the S&P 500 index.
method Three-step approach combining financial news sentiment and macroeconomic indicators.
result Proposed three-step ensemble approach significantly improves bubble prediction accuracy.

This study analyzes app reviews to understand students' behavior in the app market.

problem Extracting sentiment from growing app reviews manually is impractical.
method Used machine learning algorithms with TF-IDF for text representation and ensemble learning for evaluation.
result SVM achieved the highest accuracy (93.37%) on tri-gram + TF-IDF scheme.