Partition optimization for a random process realization to estimate its expected value

Abstract

The paper provides an analytical proof the optimal number of partitions of a non-stationary random process realization, which is necessary for estimating its expected value when using “the estimation reproduction” method for signal processing. This method allows to process signal with a limited volume of priori information about the desired signal function and statistical characteristics of the additive noise component.

Keywords

• desired signal
• noise
• signal-to-noise ratio
• expected value
• estimation
• random process