Institute of mathematics, physics and informatics of V. P. Astafiev Krasnoyarsk state pedagogical university, 7 Perensona str., Krasnoyarsk, 660049, Russia
Abstract:
Based on modified rescaled range scale computation algorithm, the technique of Hurst exponent and its characteristic time estimation is proposed. The approach of increase the accuracy and simplification automatic Hurst exponent calculation is developed. The Hurst exponent and characteristic time is calculated for power time sets of speech signals with various motor pathologies (aphasias and dysarthrias). Results is statistically analyzed, the correlation between Hurst exponent and characteristic time is estimated.