Fraud Detection with Autoencoders
Explore my project delving into the application of autoencoders for credit card fraud detection. I investigate various autoencoder variants—VAEs, sparse autoencoders, normal autoencoders, and denoising autoencoders—leveraging a Kaggle dataset to compare their efficacy in two scenarios: unsupervised learning with only normal transactions and mixed datasets with both normal and anomalous instances. Join me as I unravel the nuances of anomaly detection in financial transactions using advanced deep learning techniques.