TWITTER SENTIMENT ANALYSIS FOR OPTIMAL PORTFOLIO CONSTRUCTION

Year 2024, Volume: 20 Issue: 1, 17 - 23, 31.12.2024

Burak Kucukaslan Oktay Tas

Abstract

Purpose- This research investigates the efficacy of social media sentiment analysis in constructing alpha-generating investment portfolios. Specifically, the study examines whether Twitter-derived sentiment indicators can be leveraged to develop systematic trading strategies that generate risk-adjusted returns exceeding benchmark performance. The research aims to establish quantitative criteria for position initiation and termination based on sentiment metrics, with the ultimate objective of creating a portfolio that demonstrates significant outperformance relative to the reference index.
Methodology – The study encompasses 16 companies of the Nasdaq 100 index, selected to represent diverse market sectors while controlling for liquidity and market impact considerations. The dataset comprises 708,080 Twitter posts pertaining to the selected companies throughout the 2022 calendar year, extracted via programmatic data collection methodologies. Sentiment quantification was performed utilizing the Natural Language Toolkit (NLTK) in Python, generating normalized sentiment scores within a [-1, +1] interval. The investigation employed a sophisticated aggregation methodology to compute both daily and weekly sentiment indicators for each security, deliberately excluding neutral sentiment scores (0) to enhance signal clarity. A systematic portfolio construction framework was implemented, whereby securities were hierarchically ranked based on their aggregate sentiment scores on a weekly basis. Multiple portfolio permutations were tested, incorporating various combinations of long positions in top-ranked securities and short positions in bottom-ranked securities. Position entry and exit prices were determined using weekly opening and closing prices, respectively. Portfolio performance was evaluated through the calculation of weekly returns and cumulative performance metrics over the observation period.
Findings- The empirical results reveal that portfolios constructed exclusively with short positions demonstrated superior cumulative returns compared to long-only portfolios. This observation can be contextualized within the broader market environment, specifically the Nasdaq 100's negative 33% return in 2022. The research identified statistically significant outperformance in portfolios implementing a combined long-short strategy, with these portfolios generating positive absolute returns despite the challenging market conditions.
Conclusion- The empirical evidence substantiates the hypothesis that Twitter sentiment analysis can be effectively utilized as a signal generation mechanism for systematic portfolio construction. The results demonstrate statistically significant alpha generation capabilities, particularly when implementing a long-short strategy, suggesting potential applications for institutional investors and quantitative fund managers.

Keywords

Twitter sentiment analysis, algorithmic portfolio construction, market efficiency, natural language processing, behavioral finance

References

• Mendoza-Urdiales, R. A., Núñez-Mora, J. A., Santillán-Salgado, R. J., & Valencia-Herrera, H. (2022). Twitter Sentiment Analysis and Influence on Stock Performance Using Transfer Entropy and EGARCH Methods. MDPI, Entropy, 24, 874.
• Yang, S. Y., Mo, S. Y. K., Liu, A., & Kirilenko, A. A. (2017). Genetic programming optimization for a sentiment feedback strength-based trading strategy. Neurocomputing, 264, 29–41.
• Leow, E. K. W., Nguyen, B. P., & Chua, M. C. H. (2021). "Robo-advisor using genetic algorithm and BERT sentiments from tweets for hybrid portfolio optimization. Expert Systems with Applications, 179, 115060.
• Makrehchi, M., Shah, S., & Liao, W. (2013). Stock Prediction Using Event-based Sentiment Analysis. 2013 IEEE/WIC/ACM International Conferences on Web Intelligence (WI) and Intelligent Agent Technology (IAT).
• Ranco, G., Aleksovski, D., Caldarelli, G., Grčar, M., & Mozetič, I. (2015). The Effects of Twitter Sentiment on Stock Price Returns. PLoS ONE, 10(9).
• Oliveira, N., Cortez, P., & Areal, N. (2017). "The impact of microblogging data for stock market prediction: Using Twitter to predict returns, volatility, trading volume and survey sentiment indices. Expert Systems with Applications, 73, 125–144.
• Granholm, J., & Gustafsson, P. (2017). The Quest for the Abnormal Return: A Study of Trading Strategies Based on Twitter Sentiment. Umeå School of Business and Economics, Spring semester 2017, Degree project.
• Azar, P. D., & Lo, A. W. (2016). The Wisdom of Twitter Crowds: Predicting Stock Market Reactions to FOMC Meetings via Twitter Feeds. The Journal of Portfolio Management - Special Issue 2016.
• Chamberlain, B., Liu, Z. (F.), & Smales, L. A. (2023). Short interest and the stock market relation with news sentiment from traditional and social media sources." Aust Economics Papers, 62, 321–334.
• Sul, H. K., Dennis, A. R., & Yuan, L. (I.). (2017). Trading on Twitter: Using Social Media Sentiment to Predict Stock Returns. Decision Sciences, 48(3), 212,231.
• Yu, J. R., Chiou, W. P., Hung, C. H., Dong, W. K., & Chang, Y. H. (2022). Dynamic rebalancing portfolio models with analyses of investor sentiment. International Review of Economics & Finance, 77, 1–13.
• Tan, S. D., & Taş, O. (2021). Social Media Sentiment in International Stock Returns and Trading Activity. Journal of Behavioral Finance, 22(2), 221-234.
• Gu, C., & Kurov, A. (2020). Informational Role of Social Media: Evidence from Twitter Sentiment. Journal of Banking & Finance, 121, 105969.
• Fan, R., Talavera, O., & Tran, V. (2020). Social media, political uncertainty, and stock markets. Review of Quantitative Finance and Accounting, 55, 1137–1153.
• Kraaijeveld, O., & De Smedt, J. (2020). The predictive power of public Twitter sentiment for forecasting cryptocurrency prices. Journal of International Financial Markets, Institutions and Money, 65, 101188.
• Dang, N. C., Moreno-García, M. N., & De la Prieta, F. (2020). Sentiment Analysis Based on Deep Learning: A Comparative Study. Electronics, 9, 483.
• Broadstock, D. C., & Zhang, D. (2019). Social media sentiment and market behavior. Finance Research Letters, 30, 116-123.
• Affuso, E., & Lahtinen, K. D. (2018). Social media sentiment and market behavior. Springer-Verlag GmbH Germany, part of Springer Nature 2018.
• Hung, M. C., Hsia, P. H., Kuang, X. J., & Lin, S. K. (2023). Intelligent portfolio construction via news sentiment analysis. International Review of Economics and Finance, 89, 605–617.
• Behrendt, S., & Schmidt, A. (2018). The Twitter Myth Revisited: Intraday Investor Sentiment, Twitter Activity and Individual-Level Stock Return Volatility. Journal of Banking & Finance, 96, 355-367.
• Natural Language Toolkit (Version 3.8.1) [Computer software]. NLTK Project. https://www.nltk.org/.
• Yfinance (Version 0.2.44) [Computer software]. https://github.com/ranaroussi/yfinance