Applied pattern recognition

In this chapter, the three major subdivisions of pattern-recognition methodology are discussed (1) mapping and display, (2) clustering, and (3) classification. The procedures that must be implemented to apply pattern-recognition methods are also enumerated. Specific emphasis is placed on the application of these techniques to problems in biological and environmental analyses. 

W. O. Kwan, B. R. Kowalski, Classification of wines by applying pattern recognition to chemical composition data, J. Food Sci., 43 (1978), 1320-1323. 

Given the standard geometries of hydrogen-bonded atoms, or of helices and sheets, it is possible to apply pattern-recognition techniques to analyze the conformation of a protein. This has been done (many times) and is a useful approach to the extent that the problem is well-defined. However, in many cases helices deviate from standard geometry — especially at the ends where they can tighten up or unravel. Any investigation that requires analysis of this kind of detail is better pursued interactively. 

J. M.G., Marin, P.N., Crespo-Otero, R., Zaragoza, E.T. and Garci a-Domenech, R. (2007) Applying pattern recognition methods plus quantum and physico-chemical molecular descriptors to analyze the anabolic activity of structurally diverse steroids. J. Comput. Chem., 29, 317-333. 

The fundamental premises involved in applying pattern recognition methods to SAR studies are as follows. 

The solvents used in thin layer chromatography (TLC) can be classified by the same methods as used for stationary phases in gas chromatography. The separation characteristics of parallel TLC runs with different solvents should differ as much as possible. Massart et. al. C67, 68, 1873 applied pattern recognition methods (numerical taxonomy and cluster analysis) for an objective characterization of the solvents. 

Kuo and Jurs C164] successfully applied pattern recognition methods to the semiquantitative determination of the chlorine dosage necessary for the treatment of water in a municipal water plant. A set of 17 water-quality descriptors Ce.g. turbidity, temperature, pH, alkalinity, hardness, iron, manganese,. .. total bacteria) was used to describe a water sample. Each sample corresponds to a point in a 17-dimensional space. 

Bartzat, R., Applying pattern recognition methods and structure property correlations to determine drug carrier potential of nicotinic acid and analogize to dihydropyridine, Eur. J. Pharm. Biopharm., 59, 63, 2005. 

---