Data analysis, problems encountered

In order to set the stage for introducing different types of methods and models for analyzing multi-way data it is convenient to have an idea of what type of multivariate data analysis problems can be encountered in chemical practice. A categorization will be developed that uses the types of arrangements of the data set related to the chemical problem. 

For solving the pattern recognition problem encountered in the operation of chemical processes, the analysis of measured process data and extraction of process trends at multiple scales constitutes the feature extraction, whereas induction via decision trees is used for inductive... 

Problems encountered in HPLC analysis most often stem from a lack of knowledge of the influence of the slight variation of the experimental parameters (pH, temperature, solvent composition, flow rate, etc.). The analyst has to set up the list of parameters and their possible interactions. There are hardware parameters (e.g., flow control, temperature control, lamp current) and software parameters used to interpret and report the results from stored data. The use of factorial designs is of great help. Software such as Validation Manager, from Merck, produces, in a table for each parameter and interaction, its percentage and confidence interval as well as information to help the analyst in concluding the study. 

Data analysis involves a number of problems related to determination of the inner potential, phase shifts, multiple-scattering effects, etc. in many ways the difficulties are similar to those encountered in FEED theory.Multiple scattering may be... 

There are also a few limitations of GC-MS. The problems encountered previously by several investigators have been the poor precision and accuracy of isotope ratio data, the memory effect during the sequential analysis of samples with different isotope ratios, and the lack of suitable chelating agents. In addition, the contributions of the different isotopes of C, N, S, etc., present in the chelating agent need to be accounted for, particularly when studying the enrichments in different samples followed by tracer administration (intravenous or oral). 

Fig. 6. This schematic diagram shows waxs patterns corresponding to the crystallization of a semicrystalline polymer. One can see the problems that one can encounter in the data analysis of time-resolved data. The first frame shows the amorphous ring characteristic of a molten polymer. In the second frame the previous frame is shown as a dotted line. The amorphous ring has slightly diminished in intensity and two small peaks have started to grow. In the following frames the diffraction peaks gain in intensity while the amorphous halo keeps diminishing. Between frames 4 and 5 the diffraction peaks move to lower q-values. If one now wants to follow the intensity of the diffraction peaks one has to subtract an ever changing background and also one has to take into account the peak movements. Mathematically this is a trivial problem but to do this in an automated way with software is nontrivial.

The determination of the atomic structure of crystalline polymers by wide angle X-ray diffraction techniques presents formidable problems which cannot conveniently be treated by the conventional methods of structure analysis. Consequently, existing methods need to be modified and new approaches to the problem developed to extract the maximum amount of structural information from the available experimental data. The nature of the problems encountered will be discussed in this paper, together with a description of methods that have been utilized to overcome them. A subsequent paper will deal with the applications of these methods to the determination of the molecular structure of several crystalline polyethers. 

In solving viscoelastic stress analysis problems, assumptions on the material properties are often essential as gathering accurate time dependent data for viscoelastic properties is difficult and time consuming. Thus, one often only has properties for shear modulus, G(t) or Young s modulus, E(t), but not both. Yet of course for even the simplest assumption of a homogeneous, isotropic viscoelastic material, two independent material properties are required for solution of two or three dimensional stress analysis problems. Consequently, three assumptions relative to material properties are frequently encountered in viscoelastic stress analysis. These are incompressibility, elastic behavior in dilatation and synchronous shear and bulk moduli. Each of the common assumptions defines a particular value for either the bulk modulus or Poisson s ratio as follows. 

A data base containing C-n.m.r. parameters of monosaccharides and their glycosides, which allows computerized structure analysis of an unknown compound from its C-n.m.r. spectrum has been established, and a library of optimized geometries and conformations based on molecular mechanics calculations (MM2 CARS) has been compiled for a number of common monosaccharides. Use of the double-DANTE sequence in ID Hartmann-Hahn- and n.O.e.-spectroscopy has been recommended for overcoming some typical problems encountered in the n.m.r. analysis of carbohydrates. ... 

---