Abstract
Maximum order complexity is an important tool for measuring the nonlinearity of a pseudorandom sequence. There is a lack of tools for predicting the strength of a pseudorandom binary sequence in an effective and efficient manner. To this end, this paper proposes a neural network (NN) based model for measuring the strength of a pseudorandom binary sequence. Using the shrinking generator (SG) keystream as pseudorandom binary sequences and then calculating the unique window size (UWS) as a representation of maximum order complexity, we can demonstrate that the proposed model provides more accurate and efficient measurements than the classical method for predicting maximum order complexity. By using UWS, which is a method of pseudorandomness measurement, we can identify with higher accuracy the level of pseudorandomness of given binary sequences. As there are different randomness tests and predicting methods, we present a prediction model that has high accuracy in comparison with current methods. This method can be used to evaluate the ciphers' pseudorandom number generator (PRNG) and can also be used to evaluate the internal components by investigating their binary output sequence pseudorandomness. Our aim is to provide an application for NN pseudorandomness and in cryptanalysis in general, as well as demonstrating the models' mathematical description and implementations. Therefore, applying NN models to predict UWS utilizes two layers of pseudorandomness testing of binary sequences and is an essential cryptanalysis tool that can be extended to other fields such as pattern recognition. (C) 2020 The Authors. Published by IASE.