Calibration classical model

FIGURE 5.62. Example of calibration and validation using the classical calibration approach, (a) Initial classical model form (b) estimating concentrations (c) reconstructing the response vector (d) calculating the spectral residual (e) calculating the concentrational residual. 

Bos et al. [94] compared the performance of ANNs for modelling the Cr-Ni-Fe system in quantitative XRF spectroscopy with the classical Rasberry-Heinrich model and a previously published method applying the linear learning machine in combination with singular value decomposition. They studied whether ANNs were able to model nonlinear relationships, and also their ability to handle non-ideal and noisy data. They used more than 100 steel samples with large variations in composition to calibrate the model. ANNs were found to be robust and to perform better than the other methods. 

The classical models of spiral galaxies were constructed using rotation velocities. In contrast, the models of elliptical galaxies were found from luminosity profiles and calibrated using central velocity dispersions or motions of companion galaxies. An overview of classical methods to construct models of galaxies is given by Perek (1962). 

Alternative models may be developed to include additional covariates which are not measured with error, e.g., X = f(R,Z). The classical model is used when an attempt to measure x is made but cannot be done so due to various measurement errors. An example of this is the measurement of blood pressure. There is only one true blood pressure reading for a subject at a particular point in time, but due to minor calibration errors in the instrument, transient increases in blood pressure due to diet, etc., possible recording errors and reading errors by the nurse, etc., blood pressure is a composite variable that can vary substantially both within and between days. In this case it makes sense to try and model the observed blood pressure using Eq. (2.84). Under this model, the expected value of X is x. In regression calibration problems, the focus is on the distribution of x given X. For purposes herein, the focus will be on the classical error model. The reader is referred to Fuller (1987) and Carroll et al. (1995) for a more complete exposition of the problem. 

Stage 5 (Building the Model). This involves the classical steps of defining the conceptual model, selecting model code, calibrating the model against field data and... 

The trick in using these models is in how the eigenvectors are calculated. Note that these models base the concentration predictions on changes in the data and not on absolute absorbance measurements (which are used in all the classical models). In order to calculate the principal components analysis (PCA) model, the spectral data must change in some way. The best way to accomplish this is to vary the concentrations of the constituents of interest. As with the ILS model, there can be problems with colinearity. If the concentrations of two important constituents in the calibration samples are always present in the same ratio (for example, 2 1 of A to B, such as if dilutions were made from a single stock sample), the model will only detect one variation, not two As far as the model is concerned, all the absorbance... 

On the other hand, the scores in the S matrix are unique to each calibration spectrum, and just as a spectrum is represented by a collection of absorbances at a series of wavelengths, it can also be a series of scores for a given set of factors. Much like the classical models performed a regression of the concentration C matrix directly on the spectral absorbances in the A matrix, it is also possible to regress C against the scores S matrix. 

There are two paradigms of multivariate calibration. Classical least squares (CLS) models the instrumental response as a function of analyte concentration. 

Different calibration models, such as classical least squares and multivariate calibration approaches have been considered. 

The approach presented above is referred to as the empirical valence bond (EVB) method (Ref. 6). This approach exploits the simple physical picture of the VB model which allows for a convenient representation of the diagonal matrix elements by classical force fields and convenient incorporation of realistic solvent models in the solute Hamiltonian. A key point about the EVB method is its unique calibration using well-defined experimental information. That is, after evaluating the free-energy surface with the initial parameter a , we can use conveniently the fact that the free energy of the proton transfer reaction is given by... 

The CLS method hinges on accurately modelling the calibration spectra as a weighted sum of the spectral contributions of the individual analytes. For this to work the concentrations of all the constituents in the calibration set have to be known. The implication is that constituents not of direct interest should be modelled as well and their concentrations should be under control in the calibration experiment. Unexpected constituents, physical interferents, non-linearities of the spectral responses or interaction between the various components all invalidate the simple additive, linear model underlying controlled calibration and classical least squares estimation. 

While these models simulate the transfer of lead between many of the same physiological compartments, they use different methodologies to quantify lead exposure as well as the kinetics of lead transfer among the compartments. As described earlier, in contrast to PBPK models, classical pharmacokinetic models are calibrated to experimental data using transfer coefficients that may not have any physiological correlates. Examples of lead models that use PBPK and classical pharmacokinetic approaches are discussed in the following section, with a focus on the basis for model parameters, including age-specific blood flow rates and volumes for multiple body compartments, kinetic rate constants, tissue dosimetry,... 

The classical multivariate calibration represents the transition of common single component analysis from one dependent variable y (measured value) to m dependent variables (e.g., wavelengths or sensors) which can be simultaneously included in the calibration model. The classical linear calibration (Danzer and Currie [1998] Danzer et al. [2004]) is therefore represented by the generalized matrix relation... 

The prediction of analytical values X according to the classical indirect calibration model follows Eq. (6.75) ... 

Originally ASTRA was developed on the base of existing models that have been converted into a dynamic formulation feasible for implementation in system dynamics and allowing for closure of the feedbacks between the models. Among these models have been the macroeconomic model, ESCOT (Schade et al., 2002) and the classical four-stage transport model, SCENES (ME P, 2000). The ASTRA model then runs scenarios for the period 1990 until 2030 using the first 12 years for calibration of the model. Data for calibration stem from various sources, with the bulk of data coming from the EUROSTAT (2005) and the OECD online databases (OECD, 2005). A detailed description of ASTRA is provided by Schade (2005). 

While in classical statistics (univariate methods) modelling regards only quantitative problems (calibration), in multivariate analysis also qualitative models can be created in this case classification is performed. 

In contrast to MLR, CLS is a direct calibration method that was designed specifically for use with spectroscopic data, and whose model formula is a reflection of the classical expression of the Beer-Lambert Law for a system with mnltiple analytes ... 

For PSDs measured by GPC, we expect a greater degree of success with the simple model for retention (eq. 5). Halasz noted that the PSDs he measured were always broader than corresponding PSDs from porisimetry and capillary condensation. This is in keeping with the convolution model (eq. 7) and indicates that the PSDs measured by GPC already contain the convolution between Kqpq and the classical PSD. If this is the case, then the "effective PSDs" provided by the GPC method should be useful for the direct prediction of calibration curves. 

Infrared data in the 1575-400 cm region (1218 points/spec-trum) from LTAs from 50 coals (large data set) were used as input data to both PLS and PCR routines. This is the same spe- tral region used in the classical least-squares analysis of the small data set. Calibrations were developed for the eight ASTM ash fusion temperatures and the four major ash elements as oxides (determined by ICP-AES). The program uses PLSl models, in which only one variable at a time is modeled. Cross-validation was used to select the optimum number of factors in the model. In this technique, a subset of the data (in this case five spectra) is omitted from the calibration, but predictions are made for it. The sum-of-squares residuals are computed from those samples left out. A new subset is then omitted, the first set is included in the new calibration, and additional residual errors are tallied. This process is repeated until predictions have been made and the errors summed for all 50 samples (in this case, 10 calibrations are made). This entire set of... 

Experience in this laboratory has shown that even with careful attention to detail, determination of coal mineralogy by classical least-squares analysis of FTIR data may have several limitations. Factor analysis and related techniques have the potential to remove or lessen some of these limitations. Calibration models based on partial least-squares or principal component regression may allow prediction of useful properties or empirical behavior directly from FTIR spectra of low-temperature ashes. Wider application of these techniques to coal mineralogical studies is recommended. 

The use ofdiese classical experimental designs requires that the variables be set to prefetermined levels. Therefore, additional effon must be made to account for wiables that are not controllable. One chemometric approach is to allow these variables to vary naturally and to collect enough data to adequately modd their effect. This is the difference between the so-called natural and controllei calibration experiments (Martens and Nxs, 1989). When it is possible to mr iu e the variables, this can be done to verify that an adequate range has covered. Inverse models as discussed in Chapter 5 can then be used to implic y model their effect. (See also Appendix A.)... 

Multivariate calibration tools are used to construct models for predicting some characteristic of future samples. Chapter 5 begins with a discussion of the reasons for choosing multivariate over univariate calibration methods. The most widely used multivariate calibration tools are then presented in two categories classical and inverse methods. 

---