Objectives bounding analyses

There is a lot to unpack from this editorial. For starters, it is clear that both Churchill and Murphy identify "modern" objective methods with instrumental methods "Commenting on instrumental analysis,. . . most modern objective methods and instruments." The older wet-chemical methods are subjective—"a 15-fold increase in productivity and speed in changing from subjective methods." This usage cannot be seen to be accidental or idiosyncratic. As editor of the primary journal for analytical chemistry, Murphy is bound to be both aware and sensitive to usage in his field. The title of his column, "Modern Objectivity in Analysis," tells us that he is making a point about modern objectivity. 

Supervised methods rely on some prior training of the system with objects known to belong to the class they define. Such methods can be of the discriminant or modeling types.11 Discriminant methods split the pattern space into as many regions as the classes encompassed by the training set and establish bounds that are shared by the spaces. These methods always classify an unknown sample as a specific class. The most common discriminant methods include discriminant analysis (DA),12 the K-nearest neighbor... 

The preceding analysis shows that there is a distinct lower bound for the size of the simulation cell. Below this bound, finite size artifacts heavily bias the thermodynamics and kinetics of ion accumulation. Therefore, it seems unadvisable to design simulations using neutralizing counterions alone, especially if the objective is to understand ion-mediated conformational transitions or to understand how ions interact with a macroion. The minimal number of excess ions needed to simulate 800 mm excess monovalent salt was nearly identical 800 pairs) for all of the systems studied (32 nt A-RNA, B-DNA, and the Tar—Tar RNA kissing-loop complex), indicating a simple dependence on net macroion charge. For other macromolecular solutes or other concentrations of excess salt, a box-size... 

It is necessary to pay attention to the fact that, during recent years, the bioinorganic properties of aluminum coordination compounds have become the objects of detailed study their general aspect [354], metabolism and toxicology [355,356], complex formation with nucleozides of di- and triphosphates and nucleo-zide-bound proteins [357], and x-ray analysis of biologically important complexes... 

Mathematical optimization deals with determining values for a set of unknown variables x, X2, , x , which best satisfy (optimize) some mathematical objective quantified by a scalar function of the unknown variables, F(xi, X2, , xn). The function F is termed the objective function bounds on the variables, along with mathematical dependencies between them, are termed constraints. Constraint-based analysis of metabolic systems requires definition of the constraints acting on biochemical variables (fluxes, concentrations, enzyme activities) and determining appropriate objective functions useful in determining the behavior of metabolic systems. 

Carry out a sensitivity analysis to determine which variables have the most impact on the objective function. These are the variables that should be used as decision variables. It is also important to determine reasonable ranges for these variables and set upper and lower bound constraints. If the ranges set are too narrow, then the optimum may not be found. If they are too wide, then convergence may be difficult. 

The objective of improvement schemes is to study and validate each step of the analytical procedure of each laboratory in a collaborative manner. In the best case, each critical step of the procedure should be evaluated in an adapted exercise. The individual steps may be studied with a series of different materials in a stepwise manner. In principle, the strategy consists of starting from the most simple matrix, e.g. pure solutions and/or mixtures of compounds in solution, which are used to test the performance of the detector. The analysis of more complex matrices (e.g. raw extract, purified extract) enables testing the separation and/or clean-up steps, whereas solid samples are used to test the entire procedure. Spiked samples can be analysed to evaluate the extraction procedure, keeping in mind that a complete recovery of a spiked analyte does not mean that the same performance will be achieved with a naturally bound determinand (conversely, if a poor recovery is obtained with a spiked sample, one can assume that this procedure will not work with a natural sample). Such an approach is actually similar to the steps that should be followed when developing and validating a new method in a laboratory. 

It is of interest to know the worst-case behavior of such heuristics when applied on an NP-hard scheduling problem that is, to have an upper bound on the ratio of the value of the objective function obtained via the heuristic divided by the true optimal value. This upper bound often, but not always, lies between 1 and 2. Besides giving an indication of how bad the result of a given heuristic may turn out to be, a worst-case analysis also gives an indication of the types of problem data for which the heuristic does not work well. 

---