Paper: arXiv 2404.07224
Authors: Francisco de Arriba-Pérez, Silvia García-Méndez, José A. Regueiro-Janeiro, Francisco J. González-Castaño
Abstract
Micro-blogging sources such as the Twitter social network provide valuable real-time data for market prediction models. Investors’ opinions in this network follow the fluctuations of the stock markets and often include educated speculations on market opportunities that may have impact on the actions of other investors. In view of this, we propose a novel system to detect positive predictions in tweets, a type of financial emotions which we term “opportunities” that are akin to “anticipation” in Plutchik’s theory. Specifically, we seek a high detection precision to present a financial operator a substantial amount of such tweets while differentiating them from the rest of financial emotions in our system. We achieve it with a three-layer stacked Machine Learning classification system with sophisticated features that result from applying Natural Language Processing techniques to extract valuable linguistic information. Experimental results on a dataset that has been manually annotated with financial emotion and ticker occurrence tags demonstrate that our system yields satisfactory and competitive performance in financial opportunity detection, with precision values up to 83%. This promising outcome endorses the usability of our system to support investors’ decision making.
Complexity vs Empirical Score
- Math Complexity: 3.5/10
- Empirical Rigor: 7.0/10
- Quadrant: Street Traders — empirical and applied, lower math complexity
Why this score: The paper uses standard ML/NLP techniques without advanced mathematical derivations, but employs a structured empirical approach with a manually annotated dataset, feature engineering, and precision metrics (up to 83%), making it data-driven and practical for trading applications.
Research Flowchart
flowchart TD
A["Research Goal<br>Detect 'Opportunities'<br>in Financial Tweets"] --> B["Data<br>Manually Annotated Dataset<br>Tweets w/ Emotion & Ticker Tags"]
B --> C["Feature Engineering<br>Linguistic & Semantic Feature Extraction"]
C --> D["Stacked ML System<br>Three-Layer Classification Architecture"]
D --> E["Experimental Evaluation<br>Performance Testing"]
E --> F{"Key Findings & Outcomes"}
F --> G["Precision up to 83%<br>High Detection Specificity"]
F --> H["System Usability<br>Supports Investor Decision Making"]
Evidence Card
What the authors claim, on what data, and who has checked it. Read from the paper's own text by a language model: every quote is verbatim, nothing was re-run. How to read this card
Proposes
The paper proposes a three‑layer stacked machine‑learning classification system to detect financial “opportunity” tweets with high precision.
Source sentence
We achieve it with a three-layer stacked Machine Learning classification system with sophisticated features that result from applying Natural Language Processing techniques to extract valuable linguistic information.
Assumes 4
- A high detection precision is preferred over a large number of positive detections in financial analysis. model
Source sentence
In each of these stages, we applied a decision depth threshold to maximise opportunity detection, because in financial analysis a high precision in key categories is preferable over obtaining a large amount of positives [70].
- The size of the training dataset does not justify the use of deep‑learning techniques. data
Source sentence
Note that the size of our training datasets did not justify the application of DL techniques [71], [72].
- Repeated and spam entries are removed to ensure data quality. data
Source sentence
To ensure the quality of the data, we discarded repeated and spam entries, yielding approximately 5,000 valid tweets.
- Emotion labels are assigned by majority voting of five experts; ties are resolved randomly. data
Source sentence
At the end, the emotion tag of each entry in the dataset was computed by majority voting (in case of tie, it was selected randomly).
Data
- TxStockData S.L. tweet dataset text-news commercial
Source sentence
TxStockData S.L., a fintech company, provided us with a dataset composed of 6,000 tweets
Costs
Not applicable
Finds 4
- The system achieves precision up to 83% in detecting financial opportunity tweets. performance
Source sentence
... demonstrate that our system yields satisfactory and competitive performance in financial opportunity detection, with precision values up to 83%.
- Using all proposed features improves classifier precision by up to ~4% compared to a basic feature set. predictive-accuracy
Source sentence
Adding all features in Section III-B1 was slightly advantageous for all classifiers (up to ∼4% of improvement) except for the SVC.
- Random Forest was the best classifier for opportunity detection, followed by Gradient Descent. performance
Source sentence
Moreover, in light of the results, RF seemed the best classifier to detect opportunities, followed by the GD classifier.
- Initial single‑layer classifier achieved precision, τ1 and τ2 up to 64.33%, 70.65% and 87.31% respectively, motivating more advanced designs. performance
Source sentence
We considered these initial results (precision, τ1 and τ2 up to 64.33%, 70.65% and 87.31%, respectively) promising and they motivated us to design more advanced approaches.
Checked by others
No other paper in the archive has been found to support or contradict these findings (compared with 13 related papers). Coverage is partial, so this is neither replication nor refutation.
Code
No repository by the authors found in the paper.
Extracted 2026-11-14 · same card as JSON · A language model chose these items: check the source sentence before relying on one.
Cited References (75)
- The circumplex as a general model of the structure of emotions and personality. (2004) — link
- Using Deep Convolutional Neural Network for Emotion Detection on a Physiological Signals Dataset (AMIGOS) (2019) — link
- Unravelling Small Sample Size Problems in the Deep Learning World (2020) — link
- EmoNet: Fine-Grained Emotion Detection with Gated Recurrent Neural Networks (2017) — link
- Sequence classification for credit-card fraud detection (2018) — link
- A linked data approach to sentiment and emotion analysis of Twitter in the financial domain (2014)
- How emotions influence behavior in financial markets: a conceptual analysis and emotion-based account of buy-sell preferences (2020) — link
- The role of emotion in P2P microfinance funding: A sentiment analysis approach (2020) — link
and 67 more
- The Document Similarity Index based on the Jaccard Distance for Mail Filtering (2019) — link
- A Novel User-based Spam Review Detection (2017) — link
- N-Gram Classifier System to Filter Spam Messages from OSN User Wall (2016) — link
- Automatic Natural Language Generation Applied to Alternative and Augmentative Communication for Online Video Content Services using SimpleNLG for Spanish (2018) — link
- A library for automatic Natural Language Generation of Spanish texts (2019) — link
- Spam comments prediction using stacking with ensemble learning (2018) — link
- Stacking-Based Ensemble Learning of Self-Media Data for Marketing Intention Detection (2019) — link
- COGEMO: Cognitive-Based Emotion Detection from Patient Generated Health Reviews (2017) — link
- Implementation of wavelet packet transform and non linear analysis for emotion classification in stroke patient using brain signals (2017) — link
- Emotion Detection in E-learning Using Expectation-Maximization Deep Spatial-Temporal Inference Network (2018) — link
- A model for sentiment and emotion analysis of unstructured social media text (2018) — link
- Sentiment Analysis of Comment Texts Based on BiLSTM (2019) — link
- Sentiment analysis on stock social media for stock price movement prediction (2019) — link
- Managing extracted knowledge from big social media data for business decision making (2017) — link
- Forecasting daily stock market return using dimensionality reduction (2017) — link
- Improving stock market prediction via heterogeneous information fusion (2018) — link
- CNNpred: CNN-based stock market prediction using a diverse set of variables (2019) — link
- Forecasting the onset and course of mental illness with Twitter data (2017) — link
- Automatic identification of eyewitness messages on Twitter during disasters (2020) — link
- How mood affects the stock market: Empirical evidence from microblogs (2020) — link
- How Does Social Media Impact Bitcoin Value? A Test of the Silent Majority Hypothesis (2018) — link
- The added value of social media data in B2B customer acquisition systems: A real-life experiment (2017) — link
- Topic sentiment mining for sales performance prediction in e-commerce (2018) — link
- Predicting Vehicle Sales by Sentiment Analysis of Twitter Data and Stock Market Values (2018) — link
- Targeting Prospective Customers: Robustness of Machine-Learning Methods to Typical Data Challenges (2019) — link
- Forecasting stock market crisis events using Deep and statistical Machine Learning techniques (2018) — link
- Enhancing Aspect-Based Sentiment Analysis of Arabic Hotels’ reviews using morphological, syntactic and semantic features (2019) — link
- More than just Noise? Examining the Information Content of Stock Microblogs on Financial Markets (2018) — link
- Pachinko Prediction: A Bayesian method for event prediction from social media data (2020) — link
- Marketing analysis of wineries using social collective behavior from users’ temporal activity on Twitter (2020) — link
- Tracking and recognizing emotions in short text messages from online chatting services (2018) — link
- Incorporating stock prices and news sentiments for stock market prediction: A case of Hong Kong (2020) — link
- Beyond negative and positive: Exploring the effects of emotions in social media during the stock market crash (2020) — link
- The Limitations of Deep Learning in Adversarial Settings (2015) — link
- Generalized Isotonic Regression (2014) — link
- Using Twitter to Predict the Stock Market (2015) — link
- Trade the Tweet: Social Media Text Mining and Sparse Matrix Factorization for Stock Market Prediction (2016) — link
- Time, Temporality and Consumer Behaviour (1998) — link
- The Use of Data Mining and Neural Networks for Forecasting Stock Market Returns (2005) — link
- The Impact of Microblogging Data for Stock Market Prediction: Using Twitter to Predict Returns, Volatility, Trading Volume and Survey Sentiment Indices (2017) — link
- Textual Affect Sensing for Sociable and Expressive Online Communication (2007) — link
- Stock Market Prediction from WSJ: Text Mining via Sparse Matrix Factorization (2014) — link
- Stock Market Forecasting with Financial Micro-blog Based on Sentiment and Time Series Analysis (2017) — link
- Social Media and its Impact on Consumers Behavior (2014)
- Sentiment Classification of the Slovenian News Texts (2016) — link
- Sentiment Analysis: Mining Opinions, Sentiments, and Emotions (2015) — link
- Sentiment Analysis System Adaptation for Multilingual Processing: The Case of Tweets (2015) — link
- Score Based Financial Forecasting Method by Incorporating Different Sources of Information Flow into Integrative River Model (2016) — link
- Quantifying the Relationship Between Financial News and the Stock Market (2013) — link
- Pre-processing Online Financial Text for Sentiment Classification: A Natural Language Processing Approach (2014) — link
- Predictive Sentiment Analysis of Tweets: A Stock Market Application (2013) — link
- Predicting Crime Using Twitter and Kernel Density Estimation (2014) — link
- Natural Language Processing in Accounting, Auditing and Finance: A Synthesis of the Literature with a Roadmap for Future Research (2016) — link
- Natural Language Based Financial Forecasting: a Survey (2018) — link
- Multi-Strategy Sentiment Analysis of Consumer Reviews Based on Semantic Fuzziness (2018) — link
- Investment Decisions Based on EEG Emotion Recognition (2017) — link
- GTI-Gradiant at TASS 2015: A Hybrid Approach for Sentiment Analysis in Twitter (2015)
- Financial News Predicts Stock Market Volatility Better than Close Price (2018) — link
- Emotions in Social Psychology: Essential Readings (2001)
- Do Financial Blogs Serve an Infomediary Role in Capital Markets? (2016) — link
- Depeche Mood: a Lexicon for Emotion Analysis from Crowd Annotated News (2014) — link
- Deep Learning for Financial Sentiment Analysis on Finance News Providers (2016) — link
- Creating Emoji Lexica from Unsupervised Sentiment Analysis of their Descriptions (2018) — link
- Can Internet Search Queries Help to Predict Stock Market Volatility? (2016) — link
- Brand-Related Twitter Sentiment Analysis Using Feature Engineering and the Dynamic Architecture for Artificial Neural Networks (2016) — link
- Audio–Visual Emotion-Aware Cloud Gaming Framework (2015) — link
- Artefacts of Management Academe (2005) — link
0 of 75 cited works are in the corpus; resolved via arXiv source bundles, Semantic Scholar and OpenAlex. Corpus-wide: 2596 linked papers.