Class definition and mixture class definition by means of construction of convex hull boundaries: application to analysis for animal fat adulteration.

A visual classification technique based on the construction of convex hull boundaries in combination with a principal component analysis is described. This combined technique was evaluated in the situation in which a distinction has to be made between 2 pure animal fat classes and the corresponding mixture class. In the first instance, a principal component analysis is carried out to ensure the 2-dimensional and thus visual aspect of the technique. Convex hulls are then constructed in the 2-dimensional principal component plane to delimit the boundaries of the different classes to be defined. The effectiveness of the constructed hull boundaries in the definition of class-membership was investigated by means of the classification of different simulated test samples. The results show that, at least for the tested applications, the technique is valid, although some false positive classifications occur. The detection of outliers especially seemed to pose problems. Therefore, some propositions are made of how to refine the developed hull technique to enhance the classification results.

Investigation of the geographical distribution of female cancer patterns in Belgium using pattern recognition techniques.

Different multivariate methods have been applied to obtain an overview of the Belgian female cancer mortality distribution. The resulting maps and figures show the patterns of female cancer mortality to be strongly geographically determined. Two major trends can be derived namely along the north-south axis and along the east-west axis of the country. The cancer sites responsible for these patterns are identified.

Feasibility study concerning the use of expert systems for the development of procedures in pharmaceutical analysis.

The feasibility of using expert systems for the development of analytical procedures is investigated. A system for the computer generation of procedures to determine active drug substances in commercial formulations is proposed. It is shown that in nearly 85% of the cases investigated the present system immediately yields a correct procedure or conclusion. It is concluded that selecting methods and developing procedures with the use of expert systems is difficult but feasible.