an:06148103
Zbl 1274.68376
Brailovsky, Victor L.; Har-Even, Michael
Detecting a data set structure through the use of nonlinear projections search and optimization
EN
Kybernetika 34, No. 4, 375-380 (1998).
00316056
1998
j
68T10 62H30
cluster structure; nonlinear projection
Summary: Detecting a cluster structure is considered. This means solving either the problem of discovering a natural decomposition of data points into groups (clusters) or the problem of detecting clouds of data points of a specific form. In this paper both these problems are considered. To discover a cluster structure of a specific arrangement or a cloud of data of a specific form a class of nonlinear projections is introduced. Fitness functions that estimate to what extent a given subset of data points (in the form of the corresponding projection) represents a good solution for the first or the second problem are presented. To find a good solution one uses a search and optimization procedure in the form of evolutionary programming. The problems of cluster validity and robustness of algorithms are considered. Examples of applications are discussed.