×

zbMATH — the first resource for mathematics

Symbolic discrimination rules. (English) Zbl 0976.62061
Bock, Hans-Hermann (ed.) et al., Analysis of symbolic data. Exploratory methods for extracting statistical information from complex data. Berlin: Springer. Studies in Classification, Data Analysis, and Knowledge Organization. 244-265 (2000).
A tree-growing algorithm is proposed in order to treat explicitly probabilistic data. A population of \(n\) objects partitioned into \(m\) classes is considered. Each object is formally described by symbolic data of probabilistic assertion type. The aim is: (i) to describe in the form of a binary tree the various classes of the given partition; (ii) to build a decision rule (discrimination or classification rule) which is able to classify new objects (with unknown class membership) to one class of the given partition. Non-probabilistic Boolean descriptions can also be handled in the proposed framework by transforming to modal variables with uniform probability functions.
For the entire collection see [Zbl 1039.62501].

MSC:
62H30 Classification and discrimination; cluster analysis (statistical aspects)
62C99 Statistical decision theory
68P01 General topics in the theory of data
PDF BibTeX XML Cite