Scholarlink

<p>Credit card fraud detection is crucial for financial institutions. This study introduces a novel framework comprising a Bayesian-optimized supervised autoencoder (BSAE) for data representation and a unique ensemble method featuring a novel voting mechanism based on neural networks and Bayesian optimization. The employed ensemble method is a blender method that incorporates the Bayesian optimization method to select the best combination of models for a given task. Experiments with various single classifiers show that models trained on BSAE-transformed data outperform those trained on principal component analysis (PCA) and original data. The blending ensemble also surpasses state-of-the-art methods, achieving F1-scores of 0.8642 on the European Credit Card Fraud Detection dataset and 0.6498 on the German Credit Data.</p>

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

Credit card fraud detection using a Bayesian-optimized deep supervised autoencoder with blending ensemble learning

  • 期刊论文

  • 发表日期：

    2026年1月2日

查看全文
  • Hossein Abbasimehr,
  • Hosein Fanai

摘要

Credit card fraud detection is crucial for financial institutions. This study introduces a novel framework comprising a Bayesian-optimized supervised autoencoder (BSAE) for data representation and a unique ensemble method featuring a novel voting mechanism based on neural networks and Bayesian optimization. The employed ensemble method is a blender method that incorporates the Bayesian optimization method to select the best combination of models for a given task. Experiments with various single classifiers show that models trained on BSAE-transformed data outperform those trained on principal component analysis (PCA) and original data. The blending ensemble also surpasses state-of-the-art methods, achieving F1-scores of 0.8642 on the European Credit Card Fraud Detection dataset and 0.6498 on the German Credit Data.