Investigating correlation between cryptocurrency prices and Twitter sentiment using XLNet
Stock market prediction has been one of the most widely researched and lucrative topics over the past few decades. Researchers and computer scientists have used many different artificial intelligence techniques in order to predict the volatile ebbs and flows of the stock market. One such techniqu...
| Main Author: | |
|---|---|
| Other Authors: | |
| Format: | Final Year Project (FYP) |
| Language: | English |
| Published: |
Nanyang Technological University
2021
|
| Subjects: | |
| Online Access: | https://hdl.handle.net/10356/153210 |
| _version_ | 1824456532225425408 |
|---|---|
| author | Lee, Nicholas Qin Shan |
| author2 | Althea Liang |
| author_facet | Althea Liang Lee, Nicholas Qin Shan |
| author_sort | Lee, Nicholas Qin Shan |
| collection | NTU |
| description | Stock market prediction has been one of the most widely researched and lucrative topics over the
past few decades. Researchers and computer scientists have used many different artificial
intelligence techniques in order to predict the volatile ebbs and flows of the stock market. One
such technique used is sentiment analysis, where the correlation between how individuals are
feeling, or their sentiment, and the prices of stocks is analyzed. With the recent boom in the
cryptocurrency market, many of the same ideas are being transferred to analyze the growing
market, sentiment analysis being one of them.
This project aims to study the correlation between sentiment of individuals and potential
investors on the popular social media website Twitter, and the effect it has on 3 cryptocurrencies:
Bitcoin (BTC), Ethereum (ETH) and Litecoin (LTC). Tweets over the years 2018 to 2020 were
scraped from Twitter and cleaned, and sentiment analysis was performed on them to determine
the correlation. Lastly, a model was trained in XLNet in order to capture this relationship, so that
it could be combined with other models in order to capture richer relationships with the financial
data.
The results show that by itself, there is a decently high correlation between Twitter sentiment and
the prices of cryptocurrencies. Also, the model was able to closely capture the relationship
between the sentiment and the currencies, making it suitable for use in further, more in-depth
studies. |
| first_indexed | 2025-02-19T03:55:36Z |
| format | Final Year Project (FYP) |
| id | ntu-10356/153210 |
| institution | Nanyang Technological University |
| language | English |
| last_indexed | 2025-02-19T03:55:36Z |
| publishDate | 2021 |
| publisher | Nanyang Technological University |
| record_format | dspace |
| spelling | ntu-10356/1532102021-11-16T04:59:14Z Investigating correlation between cryptocurrency prices and Twitter sentiment using XLNet Lee, Nicholas Qin Shan Althea Liang School of Computer Science and Engineering qhliang@ntu.edu.sg Engineering::Computer science and engineering::Computing methodologies::Artificial intelligence Stock market prediction has been one of the most widely researched and lucrative topics over the past few decades. Researchers and computer scientists have used many different artificial intelligence techniques in order to predict the volatile ebbs and flows of the stock market. One such technique used is sentiment analysis, where the correlation between how individuals are feeling, or their sentiment, and the prices of stocks is analyzed. With the recent boom in the cryptocurrency market, many of the same ideas are being transferred to analyze the growing market, sentiment analysis being one of them. This project aims to study the correlation between sentiment of individuals and potential investors on the popular social media website Twitter, and the effect it has on 3 cryptocurrencies: Bitcoin (BTC), Ethereum (ETH) and Litecoin (LTC). Tweets over the years 2018 to 2020 were scraped from Twitter and cleaned, and sentiment analysis was performed on them to determine the correlation. Lastly, a model was trained in XLNet in order to capture this relationship, so that it could be combined with other models in order to capture richer relationships with the financial data. The results show that by itself, there is a decently high correlation between Twitter sentiment and the prices of cryptocurrencies. Also, the model was able to closely capture the relationship between the sentiment and the currencies, making it suitable for use in further, more in-depth studies. Bachelor of Engineering (Computer Science) 2021-11-16T04:59:13Z 2021-11-16T04:59:13Z 2021 Final Year Project (FYP) Lee, N. Q. S. (2021). Investigating correlation between cryptocurrency prices and Twitter sentiment using XLNet. Final Year Project (FYP), Nanyang Technological University, Singapore. https://hdl.handle.net/10356/153210 https://hdl.handle.net/10356/153210 en application/pdf Nanyang Technological University |
| spellingShingle | Engineering::Computer science and engineering::Computing methodologies::Artificial intelligence Lee, Nicholas Qin Shan Investigating correlation between cryptocurrency prices and Twitter sentiment using XLNet |
| title | Investigating correlation between cryptocurrency prices and Twitter sentiment using XLNet |
| title_full | Investigating correlation between cryptocurrency prices and Twitter sentiment using XLNet |
| title_fullStr | Investigating correlation between cryptocurrency prices and Twitter sentiment using XLNet |
| title_full_unstemmed | Investigating correlation between cryptocurrency prices and Twitter sentiment using XLNet |
| title_short | Investigating correlation between cryptocurrency prices and Twitter sentiment using XLNet |
| title_sort | investigating correlation between cryptocurrency prices and twitter sentiment using xlnet |
| topic | Engineering::Computer science and engineering::Computing methodologies::Artificial intelligence |
| url | https://hdl.handle.net/10356/153210 |
| work_keys_str_mv | AT leenicholasqinshan investigatingcorrelationbetweencryptocurrencypricesandtwittersentimentusingxlnet |