Comparing Models with Experimental Results
The subject of this chapter is parameter estimation, an essential part of any analytical experimentation. Such experimentation involves the following steps. Data is collected describing the variation of one component with another. A model is devised showing the proposed form of the interaction between the components, and an equation is written to describe it. Values of the parameters in the equation must then be estimated which are compatible with the data and its precision. For a model based on theory, the parameters have physical meaning. In the case of empirical models the equations are non-unique and can be modified in any way until suitable parameters are found; furthermore, the parameters have no physical meaning. In either case the equations representing the model become tenable only with the evaluation of feasible parameter values.
KeywordsObjective Function Monte Carlo Simulation Condition Number Digital Computer Sensitivity Coefficient
Unable to display preview. Download preview PDF.