The rise of cryptocurrencies like Bitcoin and Ethereum has unlocked a mountain of detailed transaction data. Unlike traditional cash, this data allows for in-depth research into detecting unusual activities, combating money laundering, and understanding user behavior. One way to analyze this vast amount of data is through temporal networks. These networks use time-based metrics and models to make sense of the complex transaction patterns. However, the scale of the data poses a challenge for traditional graph analysis methods.

To tackle this, researchers focused on temporal motifs — small, recurring patterns of transactions over time. They studied three datasets: two from Bitcoin and one from NFTs (Non-Fungible Tokens). By examining sequences of three transactions involving up to three users, they found that simply counting these motifs over all users and time can provide a skewed understanding.

Instead, they delved deeper, looking at motifs contributed by individual users. Surprisingly, they discovered a "long-tail" distribution, meaning a few key players dominate specific motif signatures. When they split the data into different time periods, they uncovered events and unusual activities that were hidden when viewed as a whole.

Another interesting finding was the completion time of these motifs, revealing the dynamics driven by both humans and algorithms. This dynamic interplay between human and automated behavior offers new insights into how cryptocurrencies operate.

Uncovering Hiddencoin: A Closer Look at Time-Based Crypto Patterns

Actions