Real Data Examples

So far questions have been answered using simulated data, which is fine if comparisons are made. The question remains How well does the method work on real data Not much work has been performed to answer this question. Some analysis of data by Qiao et al. [22] for adsorption of N2 on MCM-41 porous materials has been successfully performed [23]. MCM-41 material has been described extensively in the literature since its discovery and development [24, 25]. It is a regular uniform mesoporous material for which the pore size may be varied depending upon prepai ation. The advantage of the specific data used is that X-ray analysis of the material was performed that yielded the packing distances between pores. With an assumption about the wall thickness between the pores, the pore radius is easily calculated. 

The effect of changing 7 on the answer for the pore radius, r. The answer is the answer for Tp as a percent of the original 

Sample designator X-ray lOO (nm) Adsorption (nm) Wall thickness (nm) Desorption (nm) WaU thickness (nm) 

---

Besides the mathematical outline, the methods are applied to real data examples from chemometrics for supporting better understanding and applicability of the methods. Prerequisites and limitations for the applicability are discussed, and results from different methods are compared. 

Fattinger, K. E. and Verotta, D., A nonparametric subject-specific population method for deconvolution I. Description, internal validation, and real data examples, J. Pharmacokinet. Biopharm., 23 581-610, 1995. 

Classification of seismic horizons, denoted extrema classification, has been described, and has proved valuable for mapping automatically seismic horizons in structurally complex regions. Furthermore, a new procedure for automatic fault displacement estimation across pre-interpreted fault surfaces is designed with the extrema classification as the core methodology. The performance of the extrema classification method has been illustrated through a set of real data examples. 

These responses are used for response in the concentration metameter for the fit for the second curve. For example, the response defined by real data for curve 1 at 3 nM is 0.06. The corresponding equiactive concentration from curve 2 is given by Equation 12.6, with Response = 0.06, basal = 0, and the values of n, E nax, and EC derived from the fit... 

The data points are fit to an appropriate function (Equation 12.5). (See Figure 12.10b.) From the real data points and calculated curves, equiactive concentrations of agonist in the absence and presence of the antagonist are calculated (see Section 12.2.1). For this example, real data points for the blocked curve were used and the control concentrations calculated (control curve Emax=1.01, n = 0.9, and EC5ij = 10 pM). The equiactive concentrations are shown in Table 12.9b. 

Regarding current ab initio calculations it is probably fair to say that they are not really ab initio in every respect since they incorporate many empirical parameters. For example, a standard HF/6-31G calculation would generally be called "ab initio", but all the exponents and contraction coefficients in the basis set are selected by fitting to experimental data. Some say that this feature is one of the main reasons for the success of the Pople basis sets. Because they have been fit to real data these basis sets, not surprisingly, are good at reproducing real data. This is said to occur because the basis set incorporates systematical errors that to a large extent cancel the systematical errors in the Hartree-Fock approach. These features are of course not limited to the Pople sets. Any basis set with fixed exponent and/or contraction coefficients have at some point been adjusted to fit some data. Clearly it becomes rather difficult to demarcate sharply between so-called ab initio and semi-empirical methods.4... 

After the dominant independent variables have been brought under control, many small and poorly characterized ones remain that limit further improvement in modeling the response surface when going to full-scale production, control of experimental conditions drops behind what is possible in laboratory-scale work (e.g., temperature gradients across vessels), but this is where, in the long term, the real data is acquired. Chemistry abounds with examples of complex interactions among the many compounds found in a simple synthesis step,... 

Life is a complicated affair otherwise none of us would have had to attend school for so long. In a similar vein, the correct application of statistics has to be learned gradually through experience and guidance. Often enough, the available data does not fit simple patterns and more detective work is required than one cares for. Usually a combination of several simple tests is required to decide a case. For this reason, a series of more complex examples was assembled. The presentation closely follows the script as the authors experienced it these are real-life examples straight from the authors fields of expertise with only a few of the more nasty or confusing twists left out. 

This book focuses on statistical data evaluation, but does so in a fashion that integrates the question—plan—experiment—result—interpretation—answer cycle by offering a multitude of real-life examples and numerical simulations to show what information can, or cannot, be extracted from a given data set. This perspective covers both the daily experience of the lab supervisor and the worries of the project manager. Only the bare minimum of theory is presented, but is extensively referenced to educational articles in easily accessible journals. 

The first edition of this textbook was praised for its vast number of graphs and data that can be used as reference. The new, second edition further strengthens this attribute with a new appendix containing material property graphs for the commonly used polymers. However, the most important change implemented in this edition is the introduction of real-world examples and a variety of problems at the end of each chapter. 

With real data, a more scientifically valid approach would be to correct the nonlinearity from physical theory. In the current case, for example, a scientifically valid approach would be to convert the data to transmission mode, subtract the stray light and reconvert to absorbance the nonlinear wavelengths would have become linear again. There are, of course, several things wrong with this procedure, all of them stemming from the fact that this data was created in a specific way for a specific purpose, not necessarily to be representative of real data ... 

Similarly, many different types of functions can be used. Arden discusses, for example, the use of Chebyshev polynomials, which are based on trigonometric functions (sines and cosines). But these polynomials have a major limitation they require the data to be collected at uniform -intervals throughout the range of X, and real data will seldom meet that criterion. Therefore, since they are also by far the simplest to deal with, the most widely used approximating functions are simple polynomials they are also convenient in that they are the direct result of applying Taylor s theorem, since Taylor s theorem produces a description of a polynomial that estimates the function being reproduced. Also, as we shall see, they lead to a procedure that can be applied to data having any distribution of the X-values. 

From the authors experience not all real data sets can be transformed to constant variance using power transformations. Instrumentation imperfections in our laboratory resulted in data that had variable variances despite our attempts at transformation. The transformed chlorothalonil data set, as shown in Table III illustrates a set where the transformations attempted nearly failed to give constant variance across the response range in this case the Hartley criterion was barely satisfied. The replications at the 0.1 and 20. ng levels had excessively high variance over the other levels. An example where constant variance was easily achieved utilized data of the insecticide chlordecone (kepone) also on the electron capture detector. Table II shows that using a transformation power of 0.3 resulted in nearly constant variance. 

Integrating real-world examples throughout the text, this volume stimulates readers to consider both fundamental theory and industrial applications. More than 100 figures elucidate the concepts described in the text. Sample questions and answers are provided where appropriate, along with detailed data and discussions. Pertinent references are offered to facilitate further study. 

Further, this book contains useful data from real-world examples that explain and stimulate the reader to consider both the fundamental theory and industrial applications. The latter is expanding rapidly and every decade brings new application areas... 

We conclude this chapter by presenting several examples of deconvolution of real data. Most of these examples represent deconvolutions of data that were used as part of a spectral analysis rather than generated as deconvolution examples or tests. The examples include high-resolution grating spectra, tunable-diode-laser (TDL) spectra, a Fourier transform infrared spectrum (FTIR), laser Raman spectra, and a high-resolution y-ray spectrum. 

This rule is best explained using real data, shown in Example 4.15. 

Now that everyone has a powerful computer in their back-pack or on their desk, it is simple to fit a straight or even a curved line to data using the TrendLine feature of Excel. As an example, let us work on some real data for DDT in trout from Lake Michigan. 

---