Application of Sentiment Analysis and Machine Learning Techniques to Predict Daily Cryptocurrency Price Returns - Senior Thesis

Aug 3, 2026

Upload by Jiven

This paper examines how social media sentiment related to Bitcoin influences daily price returns for Bitcoin and six other major cryptocurrencies. Using sentiment analysis and machine learning methods, the study evaluates whether investor sentiment provides predictive value for next-day movements. Results show that Bitcoin-related social media sentiment does not significantly improve forecasting accuracy for the six assets. Models that avoid assuming a linear relationship between current and lagged returns, while incorporating sentiment scores, outperform models that enforce linearity assumptions.

LanguageEnglish

CategoryResearch & Report

FormatPDF

Pages51

Reading Time2 h 9 min

AccessFree to download

I. Introduction

II. Literature Review

1. Sentiment Analysis

2. Machine Learning

III. Data and Methodologies

1. Data

2. Methodologies

IV. Empirical Results

1. Baseline Model Results

2. Experiment Model Results

V. Conclusion

VI. Bibliography

VII. Appendix

Claremont Colleges Scholarship @ Claremont | CMC Senior Theses | CMC Student Scholarship | | --- | --- | | 2023
Application of Sentiment Analysis and Machine Learning Techniques to Predict Daily Cryptocurrency Price Returns
Edward Wu
Follow this and additional works at: [https://scholarship.claremont.edu/cmc_theses](https://scholarship.claremont.edu/cmc_theses)
Part of the Data Science Commons, Econometrics Commons, Finance Commons, Longitudinal Data Analysis and Time Series Commons, and the Statistical Models Commons | | Recommended Citation Wu, Edward, "Application of Sentiment Analysis and Machine Learning Techniques to Predict Daily Cryptocurrency Price Returns" (2023) . CMC Senior Theses. 3220. [https://scholarship.claremont.edu/cmc_theses/3220](https://scholarship.claremont.edu/cmc_theses/3220) This Open Access Senior Thesis is brought to you by Scholarship@Claremont. It has been accepted for inclusion in this collection by an authorized administrator. For more information, please contact [scholarship@cuc.claremont.edu](scholarship@cuc.claremont.edu). Claremont McKenna College Application of Sentiment Analysis and Machine Learning Techniques to Predict Daily Cryptocurrency Price Returns Submitted to Professor Michael Gelman by Edward Wu for Senior Thesis Fall 2022 12/5/22 2 3 Acknowledgments First off, I would like to express my gratitude to my thesis advisor, Professor Michael Gelman for his invaluable guidance and feedback during this semester-long journey. Without your support and patience, it would have been difficult for my thesis to come to fruition. Additionally, I would like to thank Professor Mike Izbicki for providing me with access to historical data of all geolocated tweets dating back to 2017 as well as additional guidance with debugging my code. Lastly, I would like to thank my family and friends who provided me with immense emotional support during this tough, grueling journey. 4 Abstract This paper examines the effects of social media sentiment relating to Bitcoin on the daily price returns of Bitcoin and other popular cryptocurrencies by utilizing sentiment analysis and machine learning techniques to predict daily price returns. Many investors think that social media sentiment affects cryptocurrency prices. However, the results of this paper find that social media sentiment relating to Bitcoin does not add significant predictive value to forecasting daily price returns for each of the six cryptocurrencies used for analysis and that machine learning models that do not assume linearity between the current day price return and previous daily price returns combined with previous daily sentiment scores were more accurate than machine learning models that assume linearity. Keywords: Cryptocurrency Price Prediction, Machine Learning, Sentiment Analysis, Twitter Sentiment 5 Table of Contents I. Introduction 6 II. Literature Review 7 1. Sentiment Analysis 8 2. Machine Learning 10 III. Data and Methodologies 12 1. Data 12 2. Methodologies 15 IV. Empirical Results 20 1. Baseline Model Results 20 2. Experiment Model Results 23 V. Conclusion 26 VI. Bibliography 28 VII. Appendix 30 Appendix A: Additional Baseline Model Results 30 Appendix B: Additional Experiment Model Results 31 Appendix C: Baseline Model Results Figures 32 Appendix D: Experiment Model Results Figures 41 6 I. Introduction Although there are various studies focused on predicting stock market price movements based on social media sentiment, research applying sentiment analysis and machine learning techniques to predict cryptocurrency price movements is comparatively sparse. This paper examines the common sentiment analysis and machine learning methods utilized in studies addressing the task of predicting stock prices and assesses whether these methods can be applied to forecasting cryptocurrency prices. In comparison to the stock market, the

FAQs

What does the thesis investigate regarding Bitcoin and cryptocurrency prices?

The thesis studies whether social media sentiment related to Bitcoin can predict daily price returns for Bitcoin and other popular cryptocurrencies.

How is social media sentiment incorporated into the prediction approach?

The work applies sentiment analysis techniques, including VADER, to derive daily sentiment signals from social media content and then uses them as inputs to machine learning models.

What do the empirical results conclude about the predictive value of sentiment?

The results indicate that Bitcoin-related social media sentiment does not add significant predictive value for forecasting daily price returns across the six analyzed cryptocurrencies, and that nonlinear modeling with sentiment inputs performs better than linear-assumption models.

Aug 3, 2026

Upload by Jiven