Chemometric approaches to analytical methods development and application.

摘要

In Chapter 1, a definition of chemometrics is given. The use of experimental design is demonstrated by measuring the absorbance response from a constant volume of vanadyl sulfate as a function of the volumes of hydrogen peroxide and sulfuric acid that are added.; In Chapter 2, the sequential simplex method is used to determine the mobile phase formulation of methanol, water, and acetonitrile that will result in the optimal separation of five neutral organic solutes. The chromatographic response function and the total analysis time are used in an overall desirability function to direct the progress of the sequential simplex optimization.; In Chapter 3, a four-parameter logistic model is fitted to ten immunoassay calibration data sets using non-linear least squares. The observed behavior of some of these data sets leads to recommendations for future immunoassay calibration designs.; In Chapter 4, a tutorial on principal component analysis is given and the application of principal component analysis to an analytical methods comparison problem is presented.; In Chapter 5, the geometric properties of the constant sum, constant length, and maximum value transformations are investigated. An argument is made for using absolute values in the constant sum and maximum value transformations. In general, each transformation distorts the shape and dimensionality of patterns in the data. A data set that has been closed by one of these transformations can be reopened if a vector containing the constant sums, constant lengths, or maximum values of the original objects was retained. Transformed data sets may be freely interconverted among these three transformations without the loss of information.; Finally, the mean-center transformation is considered in Chapter 6 with emphasis on the differences between the object and dimension mean-center transformations. The object mean-center transformation alters the geometrical representation of the original data in dimension space. The dimension mean-center transformation relocates the original data at the center of the dimension coordinate axes without altering the geometry of the original data. Additional studies are presented showing the use of double-mean-centering and its application to the principal component analysis of two-way chemical data.