Unveiling underlying patterns, drivers and anomalies in cryptocurrency price dynamics through feature fusion of financial indicators and sentiment fluctuations
This thesis is submitted in partial fulfillment of the requirements for the degree of Master of Science in Computer Science, 2024.
| Päätekijä: | |
|---|---|
| Muut tekijät: | |
| Aineistotyyppi: | Opinnäyte |
| Kieli: | English |
| Julkaistu: |
Brac University
2024
|
| Aiheet: | |
| Linkit: | http://hdl.handle.net/10361/24031 |
| id |
10361-24031 |
|---|---|
| record_format |
dspace |
| institution |
Brac University |
| collection |
Institutional Repository |
| language |
English |
| topic |
Logistic regression Random forest regressor Sentiment analysis Naive Bayes Gaussian Data fusion Bitcoin. Cryptocurrencies--Prices |
| spellingShingle |
Logistic regression Random forest regressor Sentiment analysis Naive Bayes Gaussian Data fusion Bitcoin. Cryptocurrencies--Prices Rabbi, Md. Nafis Unveiling underlying patterns, drivers and anomalies in cryptocurrency price dynamics through feature fusion of financial indicators and sentiment fluctuations |
| description |
This thesis is submitted in partial fulfillment of the requirements for the degree of Master of Science in Computer Science, 2024. |
| author2 |
Alam, Md. Golam Rabiul |
| author_facet |
Alam, Md. Golam Rabiul Rabbi, Md. Nafis |
| format |
Thesis |
| author |
Rabbi, Md. Nafis |
| author_sort |
Rabbi, Md. Nafis |
| title |
Unveiling underlying patterns, drivers and anomalies in cryptocurrency price dynamics through feature fusion of financial indicators and sentiment fluctuations |
| title_short |
Unveiling underlying patterns, drivers and anomalies in cryptocurrency price dynamics through feature fusion of financial indicators and sentiment fluctuations |
| title_full |
Unveiling underlying patterns, drivers and anomalies in cryptocurrency price dynamics through feature fusion of financial indicators and sentiment fluctuations |
| title_fullStr |
Unveiling underlying patterns, drivers and anomalies in cryptocurrency price dynamics through feature fusion of financial indicators and sentiment fluctuations |
| title_full_unstemmed |
Unveiling underlying patterns, drivers and anomalies in cryptocurrency price dynamics through feature fusion of financial indicators and sentiment fluctuations |
| title_sort |
unveiling underlying patterns, drivers and anomalies in cryptocurrency price dynamics through feature fusion of financial indicators and sentiment fluctuations |
| publisher |
Brac University |
| publishDate |
2024 |
| url |
http://hdl.handle.net/10361/24031 |
| work_keys_str_mv |
AT rabbimdnafis unveilingunderlyingpatternsdriversandanomaliesincryptocurrencypricedynamicsthroughfeaturefusionoffinancialindicatorsandsentimentfluctuations |
| _version_ |
1814308097792933888 |
| spelling |
10361-240312024-09-09T21:02:18Z Unveiling underlying patterns, drivers and anomalies in cryptocurrency price dynamics through feature fusion of financial indicators and sentiment fluctuations Rabbi, Md. Nafis Alam, Md. Golam Rabiul Department of Computer Science and Engineering, Brac University Logistic regression Random forest regressor Sentiment analysis Naive Bayes Gaussian Data fusion Bitcoin. Cryptocurrencies--Prices This thesis is submitted in partial fulfillment of the requirements for the degree of Master of Science in Computer Science, 2024. Cataloged from the PDF version of the thesis. Includes bibliographical references (pages 41-44). The research provides a deep exploration of cryptocurrency price dynamics by blending technical analysis, sentiment analysis, and backtesting, aiming to reveal the hidden patterns, drivers, and irregularities in their price behaviors. As the field of cryptocurrencies gains importance, characterized by extreme price volatility and sensitivity to sentiment shifts, understanding these dynamics is vital for developing effective financial models and investment strategies. Cryptocurrencies are infamous for their unpredictable nature, often influenced by market sentiment as much, if not more, than fundamental or technical indications. This study aims to bridge the gap by evaluating the effectiveness of combining sentiment analysis with traditional technical analysis to enhance predictive accuracy and investment returns. We use various predictive models, including Support Vector Machine (SVM) and Random Forest, to evaluate their performance in different scenarios. Our findings reveal that the SVM model significantly outperforms other methods when sentiment analysis is merged. Specifically, sans sentiment analysis, the Random Forest model achieves an annual return of 3.59. Nevertheless, with sentiment analysis, the SVM model generates a distinctly higher annual return of 10.112. These results underscore the crucial role of sentiment analysis in boosting the predictive power of financial models concerning cryptocurrencies. Backtesting these models offers pragmatic insights into their effectiveness. The backtesting results show that including sentiment analysis in financial models not only enhances return metrics but also improves risk management. The superior performance of the SVM model with sentiment analysis underscores the impact of market sentiment on cryptocurrency prices, indicating that investor sentiment is a potent force that should not be ignored. The implications of these findings are substantial for both academia and practice. For researchers, this study adds to the growing body of literature on financial modeling in unstable and emerging markets, like cryptocurrencies. It presents empirical evidence supporting the merging of sentiment analysis into predictive models, thereby advancing theoretical understanding and methodological approaches in the field. For practitioners, particularly investors and financial analysts, the results provide actionable insights into optimizing investment strategies. By utilizing sentiment analysis, they can develop sturdier models that better capture market movements and investor behavior, leading to improved investment outcomes. The ability to predict price movements with increased accuracy permits more effective portfolio management and risk mitigation, which are crucial in the highly volatile cryptocurrency market. Additionally, the research accentuates the importance of continuous innovation in financial modeling techniques. As the cryptocurrency market evolves, so must the methods employed to analyze and predict its behavior. The integration of sentiment analysis represents a significant leap forward in this aspect, offering a robust tool to navigate the complexities of this emerging asset class. This research highlights the value of integrating sentiment analysis into financial models for cryptocurrencies. The findings indicate that such integration not only boosts predictive accuracy but also enhances investment returns and risk management. By advancing financial modeling techniques and providing practical insights for investment strategies, this study presents a significant contribution to both academic research and practical applications in the swiftly evolving world of cryptocurrencies. Md. Nafis Rabbi M.Sc. in Computer Science 2024-09-09T06:05:40Z 2024-09-09T06:05:40Z ©2024 2024-06 Thesis ID 22366046 http://hdl.handle.net/10361/24031 en Brac University theses are protected by copyright. They may be viewed from this source for any purpose, but reproduction or distribution in any format is prohibited without written permission. 56 pages application/pdf Brac University |