Comparative performance of the classifiers for cryptosystem identification

Abstract In the present work the problem of cryptosystem identification from their cipher texts have been addressed. The supervised classification models from statistical decision theory and Artificial Neural Network have been employed for the purpose. These classification models have been validated on known data sets from UCI repository. After validation the models have been used for crypto system identification. Several feature extraction and selection techniques have been made use of for carrying out the comparative performance of the classifiers.