Application of data mining in cryptanalysis

  • Authors:
  • Pejman Khadivi;Marjan Momtazpour

  • Affiliations:
  • Department of Electrical and Computer Engineering, Isfahan University of Technology, Isfahan, Iran;Department of Electrical and Computer Engineering, Isfahan University of Technology, Isfahan, Iran

  • Venue:
  • ISCIT'09 Proceedings of the 9th international conference on Communications and information technologies
  • Year:
  • 2009

Quantified Score

Hi-index 0.00

Visualization

Abstract

Cryptography is a popular method for information hiding and reaching confidentiality in digital world. On the other hand, cryptanalysis is an interesting and useful science from different viewpoints. All the digital materials that are processed by computers or transferred in data transmission systems have some information. However, the information, embodied in the data, is entangled with known or unknown patterns and features. As an example, each certain author uses certain writing patterns and techniques which are almost unique. While cryptography algorithms hide information from the eavesdroppers, this hiding is performed through recoding. Hence, information and the related features are still remained in the encrypted output of the cryptosystem. Then, discovering the hidden features in the encrypted texts can be used to analyze the output of a cryptography algorithm. In this paper, the application of data mining in cryptanalysis is explored. It is shown that how this attack may be employed to classify the cipher texts. Also, a number of methods to improve the security of cryptography systems will be introduced. Simulation results support the arguments of the paper.