TY - GEN
T1 - Time series analysis for bitcoin transactions
T2 - 18th IEEE International Conference on Data Mining Workshops, ICDMW 2018
AU - Toyoda, Kentaro
AU - Ohtsuki, Tomoaki
AU - Mathiopoulos, P. Takis
PY - 2019/2/7
Y1 - 2019/2/7
N2 - Due to the increased popularity of Bitcoin, many researchers have analyzed how Bitcoin is being used based on the transaction history. However, the existing works analyze the transaction history in a 'static' manner and none of them analyzes transaction history 'dynamically', i.e. without taking into account the 'time variation of how Bitcoin is transferred'. The time analysis is in great demand for many practical cases, such as digital forensics tool that infers what was going on behind the scene of a fraudulent scam, and real-time inference of marketplace sales. In this paper, we propose a novel time series analysis for analyzing the history of Bitcoin transactions. In fact the main goal of our research is to detect changing points, namely anomaly detection, against a given (Bitcoin) address's transaction history. To show the effectiveness of the proposed approach, it is tested against the transaction history of Pirate@40's HYIP (High Yielding Investment Program) scheme, which raised 700,000 BTC from his investors and was charged by the Security and Exchange Commission (SEC) in 2013. It is shown that the proposed approach can successfully detect several remarkable points of Pirate@40's HYIP scheme, such as when its program's name was changed to Bitcoin Saving & Trust and when its investment rule was changed.
AB - Due to the increased popularity of Bitcoin, many researchers have analyzed how Bitcoin is being used based on the transaction history. However, the existing works analyze the transaction history in a 'static' manner and none of them analyzes transaction history 'dynamically', i.e. without taking into account the 'time variation of how Bitcoin is transferred'. The time analysis is in great demand for many practical cases, such as digital forensics tool that infers what was going on behind the scene of a fraudulent scam, and real-time inference of marketplace sales. In this paper, we propose a novel time series analysis for analyzing the history of Bitcoin transactions. In fact the main goal of our research is to detect changing points, namely anomaly detection, against a given (Bitcoin) address's transaction history. To show the effectiveness of the proposed approach, it is tested against the transaction history of Pirate@40's HYIP (High Yielding Investment Program) scheme, which raised 700,000 BTC from his investors and was charged by the Security and Exchange Commission (SEC) in 2013. It is shown that the proposed approach can successfully detect several remarkable points of Pirate@40's HYIP scheme, such as when its program's name was changed to Bitcoin Saving & Trust and when its investment rule was changed.
KW - Bitcoin
KW - Data Mining
KW - Fraud Detection
KW - Transaction Analysis
UR - http://www.scopus.com/inward/record.url?scp=85062880591&partnerID=8YFLogxK
UR - http://www.scopus.com/inward/citedby.url?scp=85062880591&partnerID=8YFLogxK
U2 - 10.1109/ICDMW.2018.00028
DO - 10.1109/ICDMW.2018.00028
M3 - Conference contribution
AN - SCOPUS:85062880591
T3 - IEEE International Conference on Data Mining Workshops, ICDMW
SP - 151
EP - 155
BT - Proceedings - 18th IEEE International Conference on Data Mining Workshops, ICDMW 2018
A2 - Yu, Jeffrey
A2 - Li, Zhenhui
A2 - Tong, Hanghang
A2 - Zhu, Feida
PB - IEEE Computer Society
Y2 - 17 November 2018 through 20 November 2018
ER -