Enhancing the Security of E-Commerce Systems Against Various Types of Attacks Using Deep Learning Model

Nowadays, hackers can exploit unprotected e-commerce networks to attack an e-commerce network. As Malware becomes increasingly prevalent, diverse and sophisticated, recent studies highlight the effectiveness of deep Convolutional Neural Networks (CNNs) in detecting malware through attack classification. The background of this is complicated by factors such as data immutability, buyer, fraud and social network attacks. To overcome these issues, our proposed Convolution-based Buffalo Optimization (CbBO) is developed. To increase the performance, traditional buffalo optimization can be combined in a hybrid way with neural networks. The process of buffalo optimization entails identifying the best options depending on herd behavior. It may be necessary to alter how the optimization procedure searches the hyperparameter space to adapt this method to CNNs. Create a fitness feature that assesses CNN’s performance, especially regarding malware detection. Moreover, this research acts as the security classifier for e-commerce systems to identify attack variants and enhance security attack detection. Additionally, a technical model is developed to depict the dynamism of the model’s constituent parts. In the context of e-commerce security, our study will present simulation results to assess the effectiveness of the proposed paradigm. The CbBO model has been implemented using the Python platform and experimental validation demonstrates its capability to achieve 96% accuracy in effectively identifying DDoS attacks.