Publikace detail

Regression-based detection and analysis of a change point

Autoři: Meloun Milan | Kupka Karel

Rok: 2008

Druh publikace: ostatní - přednáška nebo poster

Název zdroje: Sborník přednášek konference Zajištění kvality analytických výsledků

Název nakladatele: Ing. Václav Helán - 2 THETA

Strana od-do: nestránkováno

Tituly:

Jazyk	Název	Abstrakt	Klíčová slova
cze	Regresní detekce bodu ekvivalence a bodu zvratu	Aplikací nelineární regrese na regresní model titrační křivky lze stanovit bod ekvivalence či bod zvratu. Má uplatnění také u CUSUM diagramů.	Regresní detekce; bod ekvivalence; bod zvratu.
eng	Regression-based detection and analysis of a change point	In this paper we discuss statistical tools for detection of change of an analytical signal caused by unexpected change of analyte concentration. To detect change it is advantageous to use methods based on cumulative sums and especially the CUSUM control chart technique was found to be useful. In a noisy signal however, it is hard to decide when in the time series the change took place and to locate the source or possible responsibility of the change. Here, we suggested a nonlinear regression model with a change-point parameter. The advantage of the described procedure is its capability to locate the time of the change with better accuracy and with statistical confidence intervals at a given confidence level.	change of an analytical signal; a change-point parameter; a change-point parameter;

Search