Thomas Prang
Los Alamos National Laboratory
04-30-98
The meaning of structure or pattern for data which fulfills only nominal requirements will be investigated. Basic Information- and Uncertainty-measures will be discussed and a theoretical framework of four basic techniques for structure-finding introduced. I will conclude with an overview of a set of both, well known and less known methods, which will be related to each other, to the structure-finding problem and also to the introduced basic techniques. To contrast the nominal-data, unsupervised approach some fundamental methods of the much more investigated continuous, supervised domain will also be presented.