Fraud detection in Bitcoin cryptocurrency on the Blockchain platform using CNN-BiGRU methods

Despite significant technological advancements, financial fraud continues to rise, primarily due to insufficient coordination between organizations and privacy concerns that restrict access to reliable transaction data. This study proposes a blockchain-based solution that leverages both deep learning and machine learning techniques for effective fraud detection. The model constructs a graph of sequential Bitcoin blockchain transactions, from which it extracts relevant features and classifies them to identify fraudulent activity. To improve the accuracy of fraud detection, the system incorporates a Combination of CNN Bi-GRU (Bidirectional GRU) network for classification and prediction, alongside the XGBoost algorithm for feature selection.

کلیدواژه ها:

Bitcoin ، Blockchain ، Fraud detection ، CNN ، Bi-GRU