Date of Publication :4th April 2024

Abstract:In the realm of blockchain technology, maintaining the network's integrity and reliability depends critically on the identification of fraudulent transactions. Such fraudulent operations damage trust in blockchain-based solutions like cryptocurrencies in addition to posing hazards to the economy. This study uses a broad range of machine learning methods to tackle the problem of fraud detection in blockchain networks. The efficacy of different sets of algorithms like Multilayer Perceptron (MLP), Naive Bayes, AdaBoost, Decision Tree, Random Forest, and Support Vector Machine (SVM), as well as Logistic Regression, is methodically examined. Our models are trained and assessed against the "Crypto Investment Fund Directory" dataset using the "Ethereum Fraud Detection Dataset," which includes both legitimate and fraudulent transactions. The principal aim of this undertaking is to enhance the dependability of blockchain networks by facilitating the prompt identification and resolution of fraudulent acts, thereby fostering confidence and trust among relevant parties.

Reference :