Detecting Malicious Blockchain Transactions Utilizing Anomaly Detection Techniques

Abstract

"One of the largest emergences in modern technology
is the use of cryptocurrencies, which are recorded transactions in a structure called Blockchain. Of these currencies, one of the largest is called Ethereum, which is a framework built to handle the use of smart contracts, or programs that are run on the Blockchain. Since this is a newer idea, there are many security flaws that come along with the adoption of it. While most of the transactions on the Blockchain can be recorded as legitimate, some are malicious, such as those in the recent DAO attack. In this paper, we want to use anomaly detection to help detect these malicious transactions. We use sites such as etherscan.io, which records all of the transactions on the Ethereum framework, to scrape the necessary data to establish a baseline of what a legitimate transaction should look like. Our goal is to conduct research to find malicious transactions and utilize clustering technology to help us detect when these malicious transactions are taking place."