Stability Verification of Statistical Methods for Signal Separation
Valery A. Zasov · 2018
This paper presents an algorithm for verifying the stability of statistical methods for signal separation. This algorithm is based on the calculation of singular parameter intervals for a signal mixing model. A feature of the algorithm is that it considers the effects of permutation and scaling that occur in statistical algorithms for signal separation. Computer simulation methods confirm the accuracy of the algorithm under the conditions of varying signal parameters and information channels characteristics.