A Compound Algorithm Using Neural and AES for Encryption and Compare it with RSA and existing AES
Seema Rani, Harish Mittal · 2015
In this paper we introduce a improved AES Algorithm which is a combination of AES and neural net. We trained our algorithm with plaintext text and cipher text produced by AES algorithm and test our algorithm for validation and verification either it produces same kind of cipher text pattern or not. In order to modify the performance of AES, AES is trained with the help neural network tool. In order to enhance the performance, we use the various algorithms: these are Data Division, Training, Performance and Derivative. In order to calculate performance, we use Mean Squared Error method. The Performance Plots are used to represent measures of all algorithms. All these Plots provide us information about training state, regression, Error Histogram and performance measures. The overall work provides us the improved performance over AES in order to increase security.