This paper gives an introduction to multivariate calibration from a chemometrics perspective and reviews the various proposals to generalize the well-established univariate methodology to the multivariate domain. Univariate calibration leads to relatively simple models with a sound statistical underpinning. The associated uncertainty estimation and figures of merit are thoroughly covered in several official documents. However, univariate model predictions for unknown samples are only reliable if the signal is sufficiently selective for the analyte of interest. By contrast, multivariate calibration methods may produce valid predictions also from highly unselective data. A case in point is quantification from near-infrared (NIR) spectra. With the ever-increasing sophistication of analytical instruments inevitably comes a suite of multivariate calibration methods, each with its own underlying assumptions and statistical properties. As a result, uncertainty estimation and figures of merit for multivariate calibration methods has become a subject of active research, especially in the field of chemometrics.
into the original polymer a comonomer which reduces crystallizability and increases chain flexibility.
Near-infrared reflectance spectrometry Is becoming more and more popular for quantitative analysis. Its potential for qualitative analysis has been neglected, however, due to the lack of appropriate methods of treating the data that correspond to the multiple regression analysis used for quantitative calibratlons. This study describes the use of the multivariate technique called discriminant analysis; in particular the advantages of the approach of using Mahalanobis distances is investigated.The techniques covered by the term near-infrared reflectance spectrometry have developed over a number of years since the initial work of Norris and co-workers (1-3). The initial development was for the analysis of agricultural products ( 4 , 5 ) , but the applications have since broadened to include the analysis of many types of materials.A distinguishing characteristic of the technology is the calibration of the instrument through the use of multiple regression analysis (4, 6). This has led to the development of the broad range of uses to which this type of instrumentation has been put, since it allowed workable calibrations to be generated without the need for explicit corrections for often unknown and possibly incorrigible error sources. However, the nature of the calibration process has limited the use of the technique (and the accompanying technology) to quantitative analysis of materials known to be present in the sample.More recently, other mathematical approaches to the handling of newinfrared reflectance data have been tried (7, 8) but these attempts have also been directed toward obtaining quantitative results from the available technology.Qualitative analysis via the use of near-infrared reflectance technology is virtually unknown. Rose (9) has distinguished 40 different pharmaceutical raw materials using the discriminant analysis capabilities of the SAS program package. Shenk et al. (10) have used the HAT matrix approach for the more limited case of determining whether forage samples for quantitative analysis came from the same population as the calibration samples.Identification of raw materials is a critical need in the pharmaceutical industry; pharmaceutical manufacturers have to account for their manufacturing materials from the point of entry into the plant and verify that each drum of raw material is what it is supposed to be (11). Consequently there is a requirement for a rapid measurement technique that can distinguish many different types of solid powders from each other. Near-infrared reflectance is a rapid measurement technique that is currently in routine use for performing analytical measurements on just that type of sample. For quantitative analysis the mathematical technique of multiple regression analysis has been employed for extracting the necessary information from the resulting mass of data; for qualitative analysis the corresponding mathematical technique is discriminant analysis. THEORYA chemist examining a spectrum for the purpose of determining the nature of the sample giving rise to the s...
on polymerization of olefins and diolefins in suspension and emulsion, and present a number of new measurements not published to date. The subject is considered mainly from the point of view of scientific information on the mechanism of polymerization in aqueous suspensions and emulsions, but brief mention is also made of the more important disclosures in the patent literature. The new data presented in this article refer to (1) initial rates of polymerization of styrene, methyl methacrylate, vinyl acetate and acrylonitrile in aqueous suspensions and in soap emulsions as a function of catalyst concentration, temperature, and soap concentration; (2) influence of water-soluble activators, such as sodium bisulfite; (3) influence of initial size of monomer droplets on initial rates of monomer consumption; (4) study, with the aid of the electron microscope, of size of monomer droplets and polymer particles throughout polymerization; and ( 5 ) a few experiments on side reactions in the domain of higher conversions. No attempt is made in this paper to review and appraise the very large number of recent patents (from about 1930 on), which protect special procedures on the use of various promoting, regulating, or modifying ingredients. A complete digest of this practice does not exist at present, but reference may be made to the excellent chapter on emulsion polymerization in the book of Talalay and Magat (55), to the very valuable compilation of patents by Hoseh in "India Rubber World" (27), and to the enumeration of a selected number of patents in the book of Scheiber (50) on pages 210-213.
scite is a Brooklyn-based organization that helps researchers better discover and understand research articles through Smart Citations–citations that display the context of the citation and describe whether the article provides supporting or contrasting evidence. scite is used by students and researchers from around the world and is funded in part by the National Science Foundation and the National Institute on Drug Abuse of the National Institutes of Health.
hi@scite.ai
10624 S. Eastern Ave., Ste. A-614
Henderson, NV 89052, USA
Copyright © 2024 scite LLC. All rights reserved.
Made with 💙 for researchers
Part of the Research Solutions Family.