19 August 2002 9:00-13:00
Motivation ·
Objectives ·
Target Audience ·
Outline ·
Presenter ·
Formal Concept Analysis is an unsupervised learning technique for discovering conceptual structures in of data. These structures are graphically represented as conceptual hierarchies, allowing the analysis of complex structures and the discovery of dependencies within the data. Formal Concept Analysis is a conceptual clustering technique applied in data analysis, information retrieval, and knowledge discovery; and has received increasing attention in the KDD community during the last years.
Formal Concept Analysis arose twenty years ago as a theory for the formalization of the concept of 'concept'. It is based on the philosophical understanding that a concept is constituted by two parts: its extension which consists of all objects belonging to the concept, and its intension which comprises all attributes shared by those objects. This understanding - which is also reflected by the international standard ISO 704 - allows to derive all concepts from a given context (data table) and to introduce a subsumption hierarchy.
The tutorial provides an introduction into Formal Concept Analysis and discusses its applications in KDD. In order to get experience with this instrument of analysis, the participants will do practical training on given data stocks.
The tutorial aims at introducing researchers and practitioners of related fields, as well as the general ECML/PKDD audience, to Formal Concept Analysis. It surveys theory and applications of Formal Concept Analysis, focussing on its applications in KDD.
The tutorial is suitable to the general audience of both ECML and PKDD. It is of interest to theoreticians and practitioners from data analysis, machine learning, information retrieval, data mining, knowledge discovery, and the general AI audience interested in this increasing research area.
Introduction (10 min)
Formal Contexts and Concept Lattices (40
min)
Application Examples I (10 min)
Computing Concept Lattices
(20 min)
Exercises (30 min)
Conceptual Clustering (40
min)
Exercises (30 min)
FCA-Based Mining of Association Rules (20
min)
Application Examples II (20 min)
Dr. Gerd
Stumme
Institut für Angewandte Informatik und
Formale
Beschreibungsverfahren (AIFB)
Universität Karlsruhe (TH)
D-76128
Karlsruhe
Germany
Gerd Stumme is senior researcher at the Institute of Applied Informatics and Formal Description Methods (AIFB) at the University of Karlsruhe. He received his PhD from Darmstadt University of Technology, where he co-worked for several years with Rudolf Wille, the founder of Formal Concept Analysis. Gerd Stumme published over 40 papers on Formal Concept Analysis. He also chaired several conferences about Formal Concept Analysis.