Porovnání dvou metod polynomické kalibrace
| Title in English | Comparison of two methods of polynomial calibration |
|---|---|
| Authors | |
| Year of publication | 2017 |
| Type | Article in Proceedings |
| Conference | Matematika, informační technologie a aplikované vědy (MITAV 2017) |
| MU Faculty or unit | |
| Citation | |
| Field | Applied statistics, operation research |
| Keywords | calibration; transformation function; transformation curve; maximum likelihood method; approximation of Kenward Roger type |
| Description | In this contribution we focus on comparison of two methods of polynomial calibration. A method based on maximum likelihood method and a method that use linearized model with errors in variables and Kenward Roger type of approximation. Calibration is set of tasks which gives relationship between a reference device and a calibrated device (if some special conditions are fulfilled). In this contribution we suppose that relationship between the reference and calibrated device is polynomial. This relationship can be describe by the transformation function and represented by the transformation curve. Calibration process can be divided into two parts: 1) creation of calibration model and 2) application of calibration model (measuring by calibrated device). Firstly two models of calibration for both methods are mention in the contribution. Estimation of unknown parameters of transformation function is done with use of created models. Then we figure out confidence region for transformation function and describe measuring by calibrated device. In conclusion we compare both methods with use of simulation study. |
| Related projects: |