System of methods of automated cognitive linguistic analysis of speech signals with noise

For the first time, the article presents a system of methods for automated cognitive linguistic analysis of speech signals with noise, in which, unlike existing ones, the latter is represented by a sequence of phoneme and morpheme codes identified by the criterion of the minimum of functional of relative entropy, the set of values of which is formed as a result of sequential comparison of the results of automated transcribing the studied signal with the reference phonetic alphabet of the target language. The presented system of methods made it possible, in particular, to substantiate the process of phonetic coding of language units through the analytical generalization of the results of automated transcribing of speech signals, to formalize the process of estimation of the degree of phonation variability of language units within the framework of the proposed system of methods, to formalize the interpretation of the concept of cognitive linguistic analysis of speech signals with noise in the frequency space, to propose an applied use of the obtained system of methods of cognitive linguistic analysis, for purification the speech signal from Gaussian noise. The adequacy of the theoretical apparatus and the functionality of the methods presented in the article has been proven empirically.