Symbolic objects
Boolean symbolic objectsextension of symbolic objectsmodal symbolic objectssymbolic data analysissymbolic objects
Most algorithms of symbolic data analysis present in their output the description \(d\) of a class \(C\) of individuals which is obtained by using a generalization process. Let \(\Omega\) be a set of individuals, let \(\mathcal D\) be a set of descriptions of individuals and the descriptions of classes of individuals, and let \(Y\) be a mapping defined from \(\Omega\) into \(\mathcal D\). Consider a relation \(\mathcal R\) defined on \(\mathcal D\). Denote by \([d '{\mathcal R}d]\in\mathcal L\) the result of the ``comparison between \(d '\) and \(d\) by \(\mathcal R\), where \(\mathcal L=\{\text{true, false}\}\) or \({\mathcal L}=[0,1]\). If \(\mathcal L=\{\text{true, false}\}\), then \(``[d '{\mathcal R}d]=\text{true}"\) means that there is a connection between \(d\) and \(d '\). In the case where \({\mathcal L}=[0,1]\) the value \([d '{\mathcal R}d]\) measures the degree of connection between \(d\) and \(d '\) (fuzzy relations). A symbolic object \(s\) is defined as a triple \(s=(a,{\mathcal R},d)\), where \(d\) is a description, \({\mathcal R}\) is a relation between descriptions, and \(a\) is a mapping defined from \(\Omega\) in \(\mathcal L\) depending on \({\mathcal R}\) and \(d\). Basically two different kinds of symbolic objects are distinguished: Boolean symbolic objects and modal symbolic objects. A Boolean symbolic object results if we assume that \(\mathcal L=\{\text{true, false}\}\). A modal symbolic object results if \(\mathcal L\) is the unit interval. Various other types of symbolic objects result if we consider mappings a of the type: \(a(\omega)=[h_e(Y(\omega))h_v(\mathcal R)h_i(d)]\) with some suitably chosen functions \(h_e,h_v,h_i\). In practice these mappings are typically used in the following way: \(h_e\) is a filter of the extension of the symbolic object, \(h_v\) is a filter of the descriptive variables, and \(h_i\) is a filter on the descriptions \(d\). These notions are here explained in detail. They are illustrated by various examples. For the entire collection see [Zbl 1039.62501].
- Probabilist, possibilist and belief objects for pattern recognition by data analysis
- Probabilist, possibilist and belief objects for knowledge analysis
- Symbolic objects: Order structure and pyramidal clustering
- A convex combination approach for Markov switching CAPM of interval data
- Mining maximal frequent rectangles
- Usual operations with symbolic data under normal symbolic form
- Galois lattices and some symbolic data analysis algorithms
- scientific article; zbMATH DE number 1674001 (Why is no real title available?)
- scientific article; zbMATH DE number 1674009 (Why is no real title available?)
- scientific article; zbMATH DE number 1674010 (Why is no real title available?)
- Introduction à l'approche symbolique en analyse des données
- scientific article; zbMATH DE number 736962 (Why is no real title available?)
- scientific article; zbMATH DE number 2046040 (Why is no real title available?)
- scientific article; zbMATH DE number 1395880 (Why is no real title available?)
- Clustering large data sets described with discrete distributions and its application on TIMSS data set
- Mining Description Logics Concepts with Relational Concept Analysis
- Relative and Absolute Contributions to Aid Strata Interpretation
This page was built for publication: Symbolic objects
Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q2756625)
