Abstract:Blockchain has widespread applications in the financial field but has also attracted increasing cybercrimes. Recently, phishing fraud has emerged as a major threat to blockchain security, calling for the development of effective regulatory strategies. Nowadays network science has been widely used in modeling Ethereum transaction data, further introducing the network representation learning technology to analyze the transaction patterns. In this paper, we consider phishing detection as a graph classification task and propose an end-to-end Phishing Detection Graph Neural Network framework (PDGNN). Specifically, we first construct a lightweight Ethereum transaction network and extract transaction subgraphs of collected phishing accounts. Then we propose an end-to-end detection model based on Chebyshev-GCN to precisely distinguish between normal and phishing accounts. Extensive experiments on five Ethereum datasets demonstrate that our PDGNN significantly outperforms general phishing detection methods and scales well in large transaction networks.

What problem does this paper attempt to address?

The main problem that this paper attempts to solve is: **the problem of phishing fraud detection on the Ethereum platform**. Specifically, with the wide application of blockchain technology in the financial field, cybercrimes are also increasing day by day. In particular, phishing frauds targeting cryptocurrencies have become a major threat. These problems not only damage the economic interests of users but also pose challenges to the security and trustworthiness of the blockchain. ### Problem Background - **Financial Applications of Blockchain**: Blockchain technology is widely used in the financial field, such as Bitcoin, Ethereum and other cryptocurrencies. - **Harm of Phishing Fraud**: According to statistical data, in the first half of 2017 alone, more than 30,000 users suffered a total economic loss of $225 million on the Ethereum platform, and more than half of it was caused by phishing fraud. - **Limitations of Existing Methods**: Although the existing phishing fraud detection methods use graph representation learning technology to generate account feature vectors, they fail to achieve an end - to - end architecture, resulting in the inability to effectively learn task - related features and poor performance in large - scale transaction networks. ### Research Objectives To solve the above problems, this paper proposes a phishing fraud detection framework (PDGNN) based on Graph Neural Network (GNN), aiming to improve the detection effect in the following ways: 1. **Construct a Light - weight Transaction Network**: Reduce the computational complexity by sampling and simplifying Ethereum transaction data. 2. **Extract Transaction Sub - graphs**: For known phishing accounts, extract their related transaction sub - graphs. 3. **Design an End - to - End Detection Model**: Based on Chebyshev - GCN, automatically extract account transaction behavior features to distinguish between normal accounts and phishing accounts. ### Method Innovation Points - **Light - weight Network Strategy**: Construct a light - weight transaction network through second - order Breadth - First Search (BFS) and random walk algorithms. - **Sub - graph Sampling Rules**: Propose an adaptive sub - graph sampling method to ensure an appropriate sub - graph scale and improve subsequent computational efficiency. - **Chebyshev - GCN Model**: Use the Chebyshev polynomial convolutional neural network to automatically aggregate and update account features and improve classification performance. ### Experimental Verification The paper conducted extensive experiments on five Ethereum datasets, and the results show that PDGNN is significantly superior to existing methods in phishing fraud detection, especially in large - scale transaction networks. ### Summary This research provides a brand - new perspective and technical means for solving the phishing fraud problem on the Ethereum platform. By combining graph neural networks and light - weight network strategies, efficient and accurate phishing account detection is achieved.

Phishing Fraud Detection on Ethereum using Graph Neural Network

STFN: Spatio-Temporal Fusion Network to Detect Ethereum Phishing Scams

Phishing Scams Detection in Ethereum Transaction Network.

Edge-feature Modeling-based Topological Graph Neural Networks for Phishing Scams Detection on Ethereum

Ethereum Phishing Scam Detection Based on Data Augmentation Method and Hybrid Graph Neural Network Model

Graph neural network based phishing account detection in Ethereum

Ethereum Fraud Behavior Detection Based on Graph Neural Networks.

Multiscale Feature Fusion and Graph Convolutional Network for Detecting Ethereum Phishing Scams

Graph Neural Network for Ethereum Fraud Detection

Blockchain Phishing Scam Detection via Multi-channel Graph Classification

Phishing Scam Detection on Ethereum: Towards Financial Security for Blockchain Ecosystem.

DGA-GNN: Dynamic Grouping Aggregation GNN for Fraud Detection

TTAGN: Temporal Transaction Aggregation Graph Network for Ethereum Phishing Scams Detection

Detecting Phishing Scams on Ethereum Based on Transaction Records

Who Are the Phishers? Phishing Scam Detection on Ethereum Via Network Embedding

Fishing for Fraudsters: Uncovering Ethereum Phishing Gangs With Blockchain Data

Adaptive Attention-Based Graph Representation Learning to Detect Phishing Accounts on the Ethereum Blockchain

TGC: Transaction Graph Contrast Network for Ethereum Phishing Scam Detection

Ethereum Phishing Fraud Detection Based on Heterogeneous Transaction Subnets.

2DynEthNet: A Two-Dimensional Streaming Framework for Ethereum Phishing Scam Detection

TSGN: Transaction Subgraph Networks Assisting Phishing Detection in Ethereum