Novel Approach for Generating the Key of Stream Cipher System Using Random Forest Data Mining Algorithm

Key generator is part of the stream cipher system that is responsible for generating a long random sequence of binary bits key that used in ciphering and deciphering processes. Therefore, key generator is the heart of the stream cipher system. A system with traditional key generating techniques such as using the feedback shift register is vulnerable to cypher attacks which render it to be ineffective and insecure. A suggested novel method to overcome this problem is to use a Random Forest-Data Mining (RF-DM) algorithm. It incorporates using new types of linear and nonlinear functions to mix the plain text with the key in the encryption process and the cipher text with the reverse key in the decryption process. The method is successful in satisfying the secrecy goal of the stream cipher system because it uses two types of dual randomization to baffle the attacker or cryptanalysist. By mathematical logic and prove by experiments, we generate the key successfully by encrypting messages of different sizes. This proves the ablity to generated the unique key for each message based on its length without repeating the key in most cases. In addition, by combining the use of the random forest data mining technique as the randomizating principle in the ciperhing process makes it exteremely difficult for any attacker because the attacker does not have access or understanding of the not only the technique used but also the process to decipher. The result from our experiment shows that the longer the number of the message size, the longer is the length of the generated key, which determines its strength and hence, its complexity to break by brute force.