ICFCA 2019

International Conference on Formal Concept Analysis June 25 - 28 Frankfurt, Germany
© Kevin Rupp | Frankfurt UAS

How to deal with relational data in FCA? Two complementary approaches: RCA and Graph-FCA.

ICFCA 2019 is happy to announce a tutorial on relational data in Formal Concept Analysis.


Formal Concept Analysis (FCA) can be summarized by the equation: objects+attributes = concept hierarchy. In other words, whenever you dispose of a set of objects or entities described by a set of attributes, properties, predicates, characteristics, graphs, whatever the employed vocabulary and the nature of things, FCA can help you to: (1) group a maximal set of objects sharing a maximal set of attributes into a concept, and (2) hierarchically organizing the set of concepts. Its generic perspective and its strong foundation on lattice theory make FCA a Swiss knife in knowledge engineering, information retrieval and machine learning, where it can be used to build classes and conceptual classifications, to mine implication and association rules, to build ontologies or recommendation systems, or for exploratory search. FCA is used for applications in a wide range of domains such as software engineering, environment and life sciences, and social sciences.

In this tutorial, we will present main notions and tools for applying two approaches that extend the scope of FCA to (potentially cyclic) multi-relational datasets.

Relational Concept Analysis (RCA) focuses on classifying objects from several categories both through their intrinsic attributes and through their relations to objects of other categories, propagating the concept formation from one category to another. Quantifiers inspired by description logics are involved in the concept formation, allowing a large range of analyses.

Graph-FCA (GFCA) extends attributes to n-ary predicates that represent n-ary relationships between objects. The set of objects interconnected by those relationships form a knowledge graph, similar to RDF graphs or conceptual graphs. GFCA focuses on discovering closed graph patterns where each node is interpreted as a formal concept. N-ary concepts are also defined, where the extension is a relation rather than a set.


Sébastien Ferré is an associate professor in computer science at the University of Rennes 1, France, where he teaches courses on Semantic Web, data mining, functional programming, and compilers. He is the head of the SemLIS team, in the Data and Knowledge Management (DKM) departement of the IRISA laboratory. He holds a PhD in Computer Science from the University of Rennes 1 (2002), and has also been an assistant researcher at the University of Wales, Aberystwyth. He has worked on Formal Concept Analysis (FCA), logics for knowledge representation and reasoning, information retrieval and exploration, faceted search, Semantic Web, and controlled natural languages. His application domains have been personal information management, geographical information systems, bioinformatics, software engineering, and group decision support. He has developed several tools such as Sparklis, Camelis or gfca. http://people.irisa.fr/Sebastien.Ferre/ Contact: ferre at irisa dot fr

Marianne Huchard is Full Professor of computer science at the University of Montpellier, France, where she teaches courses in software engineering and knowledge engineering. She develops her research at LIRMM (Laboratory of Informatics, Robotics and Microelectronics at Montpellier). She obtained a PhD in computer Science in 1992, during which she investigated algorithmic questions connected to multiple inheritance in various object-oriented programming languages. She conducts research in two areas: in Formal Concept Analysis (FCA) and knowledge engineering, where she studies theoretical and applied aspects, and in Software Engineering, where she contributes to model-driven engineering, component-based and service-based software development, as well as migration towards software product lines. https://www.lirmm.fr/~huchard Contact: marianne.huchard at lirmm dot fr

Xavier Dolques is an engineer from ENGEES (water and environment engineering school of Strasbourg) in Strasbourg, France. He currently works on the data mining of water quality data from national rivers. He obtained a PhD in Computer Science in 2010. His works focus on the extraction of closed relational patterns in different domains such a Model Driven Engineering to extract transformation patterns or Biological data to find relations between biological states and the water physico-chemical characteristics. His main tool being Relational Concept Analysis he developed RCAExplore. https://dolques.free.fr Contact: xavier-dolques at engees dot unistra dot fr