Estimated Accuracy of Data and Intervals of Admissible Values for Parameters of Fitting Functions
Suppose we know the function f(x,(β1,…,β J ) representing a particular natural law, where β1,…,β J are unknown free parameters to be determined by experiments. The selection of f(x,(β1,…,β J ) depends on physical theory chosen to explain the natural phenomenon considered. Given a set of experimental observations (x i , y i ) of this phenomenon and estimates of their accuracy ɛ i , i = 1,…,N, let our objective be the determination of intervals of admissible values of parameters β1,…,β J that are consistent with the estimated accuracy of experimental data.
Unable to display preview. Download preview PDF.
- Box G., Jenkins G. (1970) ‘Time Series Analysis: Forecasting and Control’, Holden-Day, San Francisco.Google Scholar
- Čepar D., Radalj Z., RibariČ M., Vovk B.(l991) ‘Determination of Admissible Values of parameters from the Estimated Accuracy of Data’, Proceedings of the 16th Symposium on Operation Research, Trier, Anton Rain Verlag. (in print).Google Scholar
- Ratkowsky David A. (1990) ‘Handbook of nonlinear regression models’, Marcel Dekker.Google Scholar