scientific article; zbMATH DE number 3874431
The paper deals with some techniques of cluster analysis with the aid of fuzzy sets. Two methods are studied in detail. The first one makes use of nonlinear iteration procedure by which affinities of the elements to the classes are successively modified according to formula \(A(i,k)^{EV}/U(k)^{EG}.\) A(i,k) stands for the affinity measure of the i-th element to the k-th class, U(k) is a mass of the k-th cluster, \(U(k)=\sum_{i}U(i,k)\), the sum is taken over all the elements and U(i,k) stands for the (i,k)-th element of the fuzzy partition matrix. EV, EG are the parameters of the procedure. The idea of the second method comes from investigations of dynamics of ecosystems, namely population genetics. The analysis is performed by replicator equations resulting from a transformation of the Volterra equation, say d ln x\({}_ i/dt=\sum_{j}G_{ij}x_ j\), where \(x_ i\) is an ecological density of the species i, \(G_{ij}\) a competition coefficient for the (i,j)-th pair of species. Both approaches are verified by high dimensional data sets.
This page was built for publication:
Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q3339956)