Predicative analytics and speech biometrics

The article presents the issues of predictive analytics using different methods and algorithms for forecasting on the basis of statistical data. The relevance of the topic is determined by the necessity of developing practical systems of automatic recognition and understanding of speech signals when studying foreign languages. Issues on the analysis of speech information recognition by means of control systems are considered. Basic features of the device for the speech system recognition which should be considered at creating an algorithm for speech recognition and delivering recommendations on the improvement of pronunciation are described. An algorithm for word and phrase analysis is offered to convert the speech acoustic signal into a chain of symbols and words. The basic principles of modern speech recognition systems are analyzed. Processing of speech phrases using the proposed mathematical algorithms of signal levels will help to determine the degree of influence of individual features on the device.