Learning to Classify Blockchain Peers According to Their Behavior Sequences

Huayun Tang, Yingying Jiao, Butian Huang, Changting Lin, Shubham Goyal, Bei Wang

Abstract: Blockchain technologies have the potential to establish novel financial service infrastructures and reshape numerous fields. A blockchain is essentially a distributed ledger maintained by a set of peers (i.e., trading nodes) that do not fully trust each other. A key challenge that blockchain faces is to precisely classify the blockchain peers into categories with respect to their behavior patterns, which will not only enable deeper insights into the blockchain network but also facilitate more effective maintenance of the various peers (in private chains). In this paper, we introduce and formulate the problem of behavior pattern classification in blockchain networks and propose a novel deep-learning-based method, termed <italic xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">PeerClassifier</i> , to address the problem. To the best of our knowledge, we are the first to formally define the problem of peer behavior classification in blockchain networks. Moreover, we conduct extensive experiments to evaluate our proposed approach. Experimental results demonstrate that <italic xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">PeerClassifier</i> is significantly more effective than the existing conventional methods.