Machine Learning Methods in Identification of Encryption Methods

Cryptanalysis is the study of methods for obtaining the meaningful message from the encrypted information without access to the secret key information that is essential for decryption. One important task in cryptanalysis when only the cipher text is available is: identification of the encryption method used. The focus of this work is on identification of the encryption method. In this work, support vector machine based methods are explored for identification of encryption method from a given cipher text. This task is considered as a pattern classification task. One approach is proposed to the identification of encryption method using the support vector machines (SVM). In this approach, the cipher text is given as input to a classifier. Because of the complexity involved in the chosen encryption methods, the performance of the classifier using the cipher text is very poor. Hetero-association models built using support vector regression (SVR) are considered for generating the partially decrypted text. Different methods are considered to represent a cipher text or a partially decrypted text by a feature vector. This feature vector is given as input to an SVM based classifier.