Paper: arXiv 2402.03353
Authors: C. Sarai R. Avila
Abstract
This study investigates the relationship between tweet sentiment across diverse categories: news, company opinions, CEO opinions, competitor opinions, and stock market behavior in the biotechnology sector, with a focus on understanding the impact of social media discourse on investor sentiment and decision-making processes. We analyzed historical stock market data for ten of the largest and most influential pharmaceutical companies alongside Twitter data related to COVID-19, vaccines, the companies, and their respective CEOs. Using VADER sentiment analysis, we examined the sentiment scores of tweets and assessed their relationships with stock market performance. We employed ARIMA (AutoRegressive Integrated Moving Average) and VAR (Vector AutoRegression) models to forecast stock market performance, incorporating sentiment covariates to improve predictions. Our findings revealed a complex interplay between tweet sentiment, news, biotech companies, their CEOs, and stock market performance, emphasizing the importance of considering diverse factors when modeling and predicting stock prices. This study provides valuable insights into the influence of social media on the financial sector and lays a foundation for future research aimed at refining stock price prediction models.
Complexity vs Empirical Score
- Math Complexity: 4.5/10
- Empirical Rigor: 7.0/10
- Quadrant: Street Traders — empirical and applied, lower math complexity
Why this score: The paper applies established time-series models (ARIMA, VAR) and sentiment analysis (VADER) to real financial and social media data, demonstrating clear empirical implementation. However, the math relies on standard statistical techniques without heavy theoretical derivations or novel algorithmic complexity.
Research Flowchart
flowchart TD A["Research Goal<br>How does tweet sentiment affect biotech stock trends?"] --> B["Data Collection"] B --> C1["Twitter Data<br>COVID-19, Vaccines, Companies, CEOs"] B --> C2["Market Data<br>10 Major Pharma Stocks"] C1 --> D["Sentiment Analysis<br>VADER Model"] C2 --> E["Preprocessing &<br>Stationarity Check"] D --> F["Modeling & Forecasting"] E --> F F --> G1["ARIMA Model<br>Univariate Analysis"] F --> G2["VAR Model<br>Multivariate with Sentiment Covariates"] G1 --> H["Key Findings<br>Complex interplay between sentiment categories & stock performance"] G2 --> H H --> I["Outcome<br>Foundation for refined stock prediction models"]
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 incorporating tweet sentiment from various categories into ARIMA and VAR models to improve stock market prediction for biotech companies.
Source sentence
A key contribution of this study is the potential improvement of stock market predictions by incorporating opinions from various tweet categories into predictive models.
Assumes 4
- ARIMA model assumes that the residuals (errors) are normally distributed. model
Source sentence
In the case of ARIMA, it assumes that the residuals (errors) are normally distributed.
- Tweets collected during market hours are representative of sentiment that influences stock prices. data
Source sentence
Tweets were collected during the hours of 8:30 AM to 3:30 PM New York Time, aligning with the stock market trading hours.
- Due to limited data, no separate validation dataset is used, assuming training‑test split is sufficient. execution
Source sentence
Due to the limited amount of data, we decided not to use a separate validation dataset.
- Sentiment scores of tweets are normally distributed, facilitating time‑series modeling. statistical
Source sentence
The sentiment scores of the company, COVID-19, vaccine, CEOs, and competitors exhibit a normal distribution.
Data
- Historical stock market data for ten biotech companies market-prices unknown
Source sentence
List of companies that stock market data were collected: ● Johnson & Johnson ● Eli Lilly ● Novo Nordisk ● AbbVie ● Merck ● Pfizer ● Roche ● AstraZeneca ● Novartis ● Moderna
- Twitter data on COVID‑19, vaccines, companies and CEOs alternative public
Source sentence
To validate our hypothesis that news, companies, and CEOs affect the stock market, we collected tweets related to COVID-19, vaccines, the ten biotech companies mentioned in Section 3.1, and their respective CEOs.
Twitter / X Yahoo Finance
Costs
Not applicable
Finds 5
- Incorporating sentiment covariates into ARIMA improved stock‑price prediction accuracy. predictive-accuracy
Source sentence
We utilized the ARIMA (AutoRegressive Integrated Moving Average) model to forecast stock market performance and found that incorporating sentiment covariates into the model led to improved predictions.
- VAR models with sentiment factors also showed significant prediction improvements for some companies. predictive-accuracy
Source sentence
In some cases, incorporating the sentiment covariates significantly improved the prediction accuracy, emphasizing the importance of considering diverse factors when modeling and predicting stock prices.
- Company sentiment scores are positively correlated with competitor sentiment scores, especially for Merck. statistical
Source sentence
The most significant result was found between company sentiment scores (companyS) and competitor sentiment scores (competitorsS) for Merck (r…
- Twitter analysis revealed higher public engagement and polarized reactions to vaccine‑related news, affecting investor sentiment. other
Source sentence
While some companies and CEOs received more favorable public opinions, certain subjects like vaccines tended to provoke more polarized reactions.
- Increased volatility was observed in after‑hours trading during the study period. performance
Source sentence
volatility in stock prices during after-hours trading.
Checked by others
No other paper in the archive has been found to support or contradict these findings (compared with 34 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.