Security Enhancement for CryptocurrencyTransactions Using Machine Learning

This paper introduces an architecture for enhancing the security of cryptocurrency transactions through machine learning techniques. The process begins with importing blockchain data and conducting exploratory data analysis (EDA) to address missing values and visualize data distributions. It also involves feature selection and managing data imbalance with the Synthetic Minority Over-Sampling Technique (SMOTE). To evaluate anomaly detection, three machine learning algorithms—Linear Regression, Naive Bayes, and XGBoost—are compared for their performance. Among them, XGBoost outperforms Linear Regression and Naive Bayes, which are often preferred in other domains. In detecting typical versus anomalous transactions, XGBoost delivers superior results. This study contributes to building robust security systems that protect Bitcoin transactions from fraudulent activities, thereby increasing trust and reliability in the cryptocurrency ecosystem. The proposed architecture integrates machine learning methods with blockchain data to bolsterthe robustness of Bitcoin transactions, reinforcing the integrity of the cryptocurrency market.