Application of Random Forest Algorithm to Detect Financial Transaction Fraud Based on Machine Learning

Abstract

The increasing number of digital financial transactions through online banking, e-wallets, and e-commerce platforms has provided convenience but also introduced new risks in the form of financial fraud. As the volume and complexity of transaction data continue to grow, fraud patterns become more diverse and increasingly difficult to detect manually. One of the main challenges in fraud detection is the limitation of conventional systems that are not able to identify suspicious transactions quickly and accurately. This issue occurs due to the imbalance between normal transaction data and fraudulent transaction data, as well as continuously evolving fraud patterns. As a result, manual detection processes become difficult and may lead to financial losses for financial institutions and service users. This study proposes the application of the Random Forest algorithm based on Machine Learning to detect financial transactions that are potentially fraudulent. The Random Forest algorithm is selected because it can handle large datasets, reduce the risk of overfitting, and provide stable classification performance. The objective of this research is to detect financial transaction fraud using the Random Forest method. The expected outcome is a model that can effectively identify suspicious transactions with minimal errors and improve the security of digital financial transactions.