PASCAL - Pattern Analysis, Statistical Modelling and Computational Learning

A Formal Concept Analysis framework to model functional dependencies.
Jaume Baixeries
In: Mathematical Methods for Learning 2004, 21-24 June 2004, Como, Italy.


Formal Concept Analysis (FCA) has been proved to be a valuable tool to represent the knowlegde contained in a database and, in general, to calculate information within it, (for instance, implication rules). FCA is primaly based on a closure operator, which allows to derive (among other information) a concept lattice for a given set of data. In this paper we define a new closure operator for multivalued datasets, which summarizes and formalizes previous work and offers a single and sound theoretical FCA framework to derive functional dependencies. We also prove that this operator generalizes a previous one that is used to find implication rules in binary databases.

EPrint Type:Conference or Workshop Item (Paper)
Project Keyword:Project Keyword UNSPECIFIED
Subjects:Theory & Algorithms
ID Code:168
Deposited By:Jaume Baixeries
Deposited On:03 June 2004