Automatic construction and refinement of a class hierarchy over semistructured data

Published 4 August 2001

Nathalie Pernelle, Marie‐Christine Rousset, Véronique Ventos

Citations13

TL;DR

An approach based on the use of two languages of description of classes for the automatic clustering of semistructured data, which has been implemented and experimented on real data in the setting of the GAEL project 1 which aims at building exible electronic catalogs organized as a hierarchy of classes of products.

Abstract

this paper, we present an approach based on the use of two languages of description of classes for the automatic clustering of semistructured data. The first language of classes has a high power of abstraction and guides the construction of a lattice of classes covering the whole set of the data. The second language of classes, more expressive and more precise, is the basis for the refinement of a part of the lattice that the user wants to focus on. Our approach has been implemented and experimented on real data in the setting of the GAEL project which aims at building flexible electronic catalogs organized as a hierarchy of classes of products. Our experiments have been conducted on real data coming from the C/Net (http://www.cnet.com) electronic catalog of computer products

Keywords

Computer Science

Automatic subspace clustering of high dimensional data for data mining applications

2,386 Citations1998Rakesh Agrawal, Johannes Gehrke +2 more

CLIQUE is presented, a clustering algorithm that satisfies each of these requirements of data mining applications including the ability to find clusters embedded in subspaces of high dimensional data, scalability, end-user comprehensibility of the results, non-presumption of any canonical data distribution, and insensitivity to the order of input records.

Machine LearningLearning from Observation: Conceptual Clustering

535 Citations1983Ryszard S. Michalski, Robert E. Stepp

Traditional techniques for this purpose, developed in cluster analysis and numerical taxonomy, are often inadequate because they arrange objects into classes solely on the basis of a numerical measure of object similarity.

arXiv (Cornell University)Top-down induction of clustering trees

398 Citations2000Hendrik Blockeel, Luc De Raedt +1 more

View PDF

Shift of bias for inductive concept learning

288 Citations1984Paul E. Utgoff

It is shown that search for appropriate bias is itself a major part of the learning task, and that mechanical procedures for conducting a well-directed search for an appropriate bias can be created.

Computing Least Common Subsumers in Description Logics with Existential Restrictions

175 Citations1999Franz Baader, Ralf Küsters +1 more

Principles of Knowledge Representation and ReasoningLearning the CLASSIC Description Logic: Theoretical and Experimental Results

157 Citations1994William W. Cohen, Haym Hirsh

Previous formal learnability results on simple description logics are extended to C-CLASSIC, a description logic expressive enough to be practically useful, and two extensions of a learning algorithm suggested by the formal analysis are experimentally evaluated.

Machine LearningA Polynomial Approach to the Constructive Induction of Structural Knowledge

113 Citations1994Jörg-Uwe Kietz, Katharina Morik

KLUSTER builds the most specific generalization and a most general discrimination in polynomial time and embeds these concept learning problems into the overall task of learning a hierarchy of concepts.

View PDF

European Conference on Artificial IntelligenceConceptual clustering in a first order logic representation

64 Citations1992Gilles Bisson

The Conceptual Clustering system KBG structures information from a set of observations and a domain theory into a directed graph of concepts, generated by an iterative use of clustering and generalization operators, both guided by similarity measures.

Lecture notes in computer scienceExtending K-Means Clustering to First-Order Representations

32 Citations2000Mathias Kirsten, Stefan Wrobel

An in-depth evaluation of two approaches of extending k-means clustering to work on first-order representations shows that in this case indeed the k-medoids approach is a viable and fast alternative to existing agglomerative or top-down clustering approaches even for a small-scale dataset, while k-prototypes exhibited a number of deficiencies.

TDIS : an Algebraic Formalization.

27 Citations1993Jean‐Gabriel Ganascia

This paper introduces a formal model to describe TDIS, based on the use of lattice theory and, more precisely, on theUse of a Galois connection, that justifies the extensive use of attribute-value representation in the case of classical TDIS and generalizes this representation to products of hierarchical posets.