Anomaly Detection in Bitcoin Blockchain: Exploring Trends and Algorithms

Abstract

The implementation of blockchain technology has become prevalent across several businesses, although its initial design was primarily intended for cryptocurrencies. The preservation of security and integrity inside blockchain networks is of utmost importance, and the process of anomaly detection serves as a crucial component in the identification and mitigation of potential threats. This research study aims to examine the patterns within the Google BigQuery Bitcoin Blockchain Dataset and utilize unsupervised anomaly detection techniques to locate anomalies within the dataset. The analysis conducted provides valuable insights into the temporal patterns exhibited by significant components of blockchain technology, including Transactions, Blocks, and Output Satoshis. Following this, we proceed to assess the effectiveness of the LSTM and Isolation Forest methods in recognising anomalous instances. The results of this study offer a significant basis for improving the security of blockchain systems by implementing anomaly detection techniques.