What is chemometrics

The term chemometrics was coined several decades ago to describe a new way of analyzing chemical data, in which elements of both statistical and chemical thinking are combined.1 Since then, chemometrics has developed into a legitimate technical field of its own, and is rapidly growing in popularity within a wide range of chemical disciplines. 

There are probably as many definitions for chemometrics as there are those who claim to practice it. However, there appear to be three elements that are consistently used in historical applications of chemometrics  

With this in mind, I ask the reader to accept my humble definition of chemometrics the application of multivariate, empirical modeling methods to chemical data  

The empirical modeling element indicates an increased emphasis on data-driven rather than theory-driven modeling of data. This is not to say that appropriate theories and prior chemical knowledge are ignored in chemometrics, but that they are not relied upon completely to model the data. In fact, when one builds a chemometric calibration model for a process analyzer, one is likely to use prior knowledge or theoretical relations of some sort regarding the chemistry of the sample or the physics of the analyzer. For example, in process analytical chemistry (PAC) applications involving absorption spectroscopy, the Beer s Law relation of absorbance vs. concentration is often assumed to be true and in reflectance spectroscopy, the Kubelka-Munk or log(l/P) relations are assumed to be true. 

The multivariate element of chemometrics indicates that more than one response variable of the analyzer is used to build a model. This is often done out of necessity 

---

First of all, what is Chemometrics According to the definition of the Chemometrics Society, it is the chemical discipline that uses mathematical and statistical methods to design or select optimal procedures and experiments, and to provide maximum chemical information by analyzing chemical data. ... 

Every scientist has designed experiments. So what is there left for us to say about that topic that chemometrics/statistics can shed some light on Well, quite a bit actually, since not all experiments are designed equally, but some are definitely more equal than others (to steal a paraphrase). Another way to say it is that every experiment is a designed experiment, but some designs are better than others. 

This leads us to the other hand, which, it should be obvious, is that we feel that Chemometrics should be considered a subfield of Statistics, for the reasons given above. Questions currently plaguing us, such as How many MLR/PCA/PLS factors should I use in my model , Can I transfer my calibration model (or more importantly and fundamentally How can I tell if I can transfer my calibration model ), may never be answered in a completely rigorous and satisfactory fashion, but certainly improvements in the current state of knowledge should be attainable, with attendant improvements in the answers to such questions. New questions may arise which only fundamental statistical/probabilistic considerations may answer one that has recently come to our attention is, What is the best way to create a qualitative (i.e., identification) model, if there may be errors in the classifications of the samples used for training the algorithm ... 

This seems to be a good stopping point. The title of this chapter is Chemometrics in Spectroscopy and for the past several chapters we have departed somewhat from that general topic to discuss in some detail the very specialized question of noise in spectra. While not outside the range of interest covered by the chapter s intent, it is somewhat near the edges of what might be considered the mainstream purview of the chapter, and it is time to return to a more mainstream discussion, or at least one closer to the center of the topic. 

This definition is convenient because it allows us to then jump directly to what is arguably the simplest Chemometric technique in use, and consider that as the prototype for all chemometric methods that technique is multiple regression analysis. Written out in matrix notation, multiple regression analysis takes the form of a relatively simple matrix equation ... 

The second critical fact that comes from equation 70-20 can be seen when you look at the Chemometric cross-product matrices used for calibrations (least-squares regression, for example, as we discussed in [1]). What is this cross-product matrix that is often so blithely written in matrix notation as ATA as we saw in our previous chapter Let us write one out (for a two-variable case like the one we are considering) and see ... 

This is where we see the convergence of Statistics and Chemometrics. The cross-product matrix, which appears so often in Chemometric calculations and is so casually used in Chemometrics, thus has a very close and fundamental connection to what is one of the most basic operations of Statistics, much though some Chemometricians try to deny any connection. That relationship is that the sums of squares and cross-products in the (as per the Chemometric development of equation 70-10) cross-product matrix equals the sum of squares of the original data (as per the Statistics of equation 70-20). These relationships are not approximations, and not within statistical variation , but, as we have shown, are mathematically (algebraically) exact quantities. 

Before describing the six habits, it is important to define what is meant by the term chemometrics. A general definition is the use of statistical and mathematical techniques to analyze chemical data. In this book, we prefer the broader definition of chemometrics as the entire process whereby data (e.g., numbers in a table) are transformed into information used for decision making. ... 

Bro, R. (2003), Multivariate calibration What is in chemometrics for the analytical chemist Anal. Chim. Acta, 500,185-194. 

There are analytical techniques such as the based on sensors arrays (electronic nose and electronic tongue) that cannot be imaginable without the help of chemometrics. Thus, one could celebrate the triumph of chemometrics...so what is wrong (Pretsch Wilkin, 2006). The answer could be supported by a very well-known quotation attributed to the british politician D lsraeli (Defernez Kemsley, 1997) but modified by us as follows "There are three kinds of lies Lies, damned lies and chemometrics". Here we have changed the original word "statistics" into "chemometrics" in order to point out the suspicion towards some chemometric techniques even within the scientific community. There is no doubt that unwarranted reliance on poorly understood chemometric methods is responsible for such as suspicion. 

However. .. This book is on chemometrics and what has chemometrics to do with cancer chemotherapy Well... understanding how these two different areas can be related to one another is the purpose of this chapter. You just must keep on reading this chapter and you will see the many ways chemometrics can be employed to investigate the "behavior" molecules exhibit considering anticancer activity and to make predictions about drugs that were not tested yet. The potential application of chemometrics to analytical data arising from problems in biology and medicine is enormous and, in fact, the applications of chemometrics have diversified substantially over the last few years (Brereton, 2007 2009). At the end of the chapter you will note that, as in many areas of research, chemometrics plays an important role in medicinal chemistry, fortunately. 

---