AHSWN Home · Issue Contents · Forthcoming Papers

Blockchain-Driven Fraud Detection in E-Commerce Transactions Using IoT, Wireless Sensor Networks, and PFILIS
Jyothi Bobba, Naresh Kumar Reddy Panga, Rajeswaran Ayyadurai, Qaisar Abbas, Karthikeyan Parthasarathy and Roseline Oluwaseun Ogundokun

Detection of fraud in the E-Commerce industry is crucial for an efficient transaction. With the integration of Internet of Things (IoT) devices supported by Wireless Sensor Networks (WSN), transaction monitoring and real-time data acquisition have become more secure and intelligent. None of the prevailing works considered detecting various types of fraud via Internet of Things (IoT) in E-commerce. Therefore, fraud classification using the Pruned Fuzzy Incremental Learning Interference System (PFILIS) for banking sector in cloud is proposed. First, the user transaction initiation is carried out and here, the level 1 authentication is done using the Universally Unique Identifier (UUI). Next, the level 2 authentication is performed using the Role-based Reinforcement Learning Access Control (RoleRLAC) to prevent attackers from accessing the account. Then, to detect fraud during transactions, the Fraud Detection Model (FDM) is trained. Here, the fraud detection data is collected and pre-processed. Further, the features are extracted and the data augmentation is done. Next, the transaction patterns are identified and, the risk assessment is carried out using the Weighted Scoring System (WSS). For low risk the transaction continues and if medium or high level risk is identified, then the fraud detection is done using the Feature Importance Convolutional Neural Networks (FICNN). After that, the type of fraud is classified by utilizing PFILIS. These details are stored in blockchain. Finally, the alert is notified to the bank and the account holder. Thus, fraud during the bank transaction is identified with an accuracy of 98.5%, and shows better performance than existing models in E-commerce cyber security.

Keywords: cyber security, e-commerce, Internet of Things (IoT), Hidden Markov Model (HMM), banking sector, data balance, Blockchain, and fraud detection during transaction, sensor network