A1 - Kowalczyk, Adam
A1 - Szlachta, Anna
A1 - Hanus, Robert
A1 - Chorzępa, Rafał
PB - Polish Academy of Sciences Committee on Metrology and Scientific Instrumentation
N2 - Autocorrelation of signals and measurement data makes it difficult to estimate their statistical characteristics. However, the scope of usefulness of autocorrelation functions for statistical description of signal relation is narrowed down to linear processing models. The use of the conditional expected value opens new possibilities in the description of interdependence of stochastic signals for linear and non-linear models. It is described with relatively simple mathematical models with corresponding simple algorithms of their practical implementation. The paper presents a practical model of exponential autocorrelation of measurement data and a theoretical analysis of its impact on the process of conditional averaging of data. Optimization conditions of the process were determined to decrease the variance of a characteristic of the conditional expected value. The obtained theoretical relations were compared with some examples of the experimental results.
L1 - http://sd.czasopisma.pan.pl/Content/107345/PDF/121.pdf
VL - vol. 24
L2 - http://sd.czasopisma.pan.pl/Content/107345
PY - 2017
IS - No 1
KW - conditional averaging
KW - conditional expected value
KW - auto-correlated data
KW - random signals
T1 - Estimation of Conditional Expected Value for Exponentially Autocorrelated Data
DA - 2017.03.30
UR - http://sd.czasopisma.pan.pl/dlibra/publication/edition/107345
DOI - 10.1515/mms-2017-0005