vol. 3, nr. 1 (2004)
|
Uncertainties of lsm approximation using transformed function in modelling agricultural processes |
|
|
Leszek Kuchar |
|
| (get PDF) |
|
|
Department of Mathematics, University of Agriculture, ul. Grunwaldzka 53, 50-357 Wrocław |
|
|
vol. 3 (2004), nr. 1,
pp. 99-105
|
|
|
abstract:
Least Square Method (LSM) as the most often used procedure in approximation fits function to the data by minimizing squared residuals. The above problem is equivalent to solving the system of equations represented by the model. It often happens, when non-linear model is selected, non-linear equation system is replaced by linear system as a consequence of function linearization and problem simplification. In this case function transformation determine different criterion of fitting, and the method estimates a different model. In the paper, uncertainties of Least Square Method (LSM) approximation and example of data fitting using linearization of exponential function are shown.
|
|
keywords:
least square method, data approximation, linear and non-linear fitting
|
|
original in:
Polish
|
|
|
|
|