Financial Fraud Detection Based on Ensemble Machine Learning

Machine learning plays an important role in credit card fraud detection. Credit card fraud detection is a huge challenge today, mainly because it has two problems – changes in fraud and normal behavior and the highly skewed dataset used which also means the ratio of fraud to normal transactions is very small. To solve the problem of unbalanced classification of data sets, this paper explores a combination of the Synthetic Minority Oversampling Technique (SMOTE) algorithm and random under-sampling for data enhancement, while using Principal Component Analysis (PCA) dimensionality reduction and removing extreme outliers in the influencing factors related to classification to reduce data redundancy. In this paper, we compare eight different machine learning algorithms and select three machine learning algorithms with better prediction results for integration learning, and find that ensemble learning can improve recall by 2%, effectively improving the accuracy of financial fraud detection.