Prediction and analysis of illegal accounts on Ethereum based on Catboost algorithm

Abstract

Increasingly frequent illegal transactions hinder the security of Ethereum transactions, and the anonymity of electronic money makes it difficult to track and analyze problems. In this paper, the transaction data of the Ethereum trading platform is used as the data source, and the marked illegal account and the unmarked normal account data set are used as the training set. Based on the CatBoost algorithm, the overall prediction of the various types of illegal accounts is made. The process adopts multiple cross-validation, the accuracy of the established algorithm model prediction reached 94.07%, and the evaluation metric of the area under the curve of the receiver reached 0.9846. The proposed scheme accurately predicts illegal behaviors on the Ethereum trading platform and effectively improves the blockchain-based trading environment.