Machine Learning Based Secured Data Transmission For Banking Application

Abstract

Security in the Internet of Things emphasizes on securing the Internet-enabled gadgets that link to wireless networks.  IoT Safety attempts to safeguard IoT devices and systems against cybercrime and it is considered to be a vital security element linked to the Internet of Things. Conversely, banking applications are progressively being supervised for their failure to provide a sufficient degree of customer support and to protect themselves against and respond to cyber-attacks. One of the main factors for this is the vulnerability of fintech systems and networks to malfunctioning. Therefore, wireless networks covering these IoT products are extremely unprotected. IoT is a lightweight system and it is optimal when using lightweight and energy-efficient cryptography for protection. Deep learning is an efficient technique to analyze threats and respond to attacks and security incidents. So this work addresses both security and energy efficiency in IoT using two novel techniques carried out through deep learning. This work contributes to the most innovative way of saving energy in IoT devices through decreasing the use of energy-expensive ‘1’ values in the interface of Dynamic RAM. This can be done by using Base + XOR encoding of data during data transmission. Also, the security of data is incorporated using chaotic XOR encryption (CXE) algorithm which is proved to perform faster and stronger encryption using XOR operation. Using Conditional Generative Adversarial Network (CGAN) based deep learning technique, the Base + XOR encoding technique and CXE are trained well in the banking application. The data generation in CGAN is carried out based on criteria produced using generator model. This work is proved to be consuming less energy, less data transmission time, and provides more security when compared to the existing systems.