Research of Accuracy Cycles Identification Methods in Symbolic Sequences with Random Noises

Keywords – periodicity, symbolic sequence, suffix trees, Fourier transformation, dynamic programming, genetic algorithm, noise in sequences. The problem of finding repeated patterns in symbolic sequences is widespread in forecasting of any type sequences that depend on time. The consequence of a common task is a considerable amount of different approaches that solves this problem.             The object of study are algorithms that implement the base approaches of the task - periodicity detection in symbolic sequence with the presence of random noises that are an inevitable constituent of real world data. The aim of the research is an algorithm analysis of finding cycle periodicity in symbolic sequences. The research requires the following points: the feature exploration, detection, and recommendation providing for existing approaches to solve the mentioned problem in a context of noisy data.   The paper presents the results of experiments on artificially generated data that covers the area of application of all studied algorithms.