Finally, we investigated autoencoder studies performed on the Credit Card Fraud Detection Dataset published by Kaggle. The relevant works are described in the following two paragraphs. Using both a plain autoencoder algorithm and a Logistic Regression algorithm, Al-Shabi [34] evaluated balanced and ...
Our approach detects the credit card transaction as genuine or anomalous. ii. To implement the model using stratified cross validation to tackle the issues of the imbalanced dataset and overfitting. Grid search hyper parameter tuning is applied to improve anomaly detection performance. We analyze the...
The TabNet network structure is further improved, and the TabNet-Stacking model, consisting of Stacking ensemble learning, is applied to classify and predict customer data from the dataset, enhancing the credit risk management and prevention levels. 4.1. Data Set Processing The data studied in this...