A Declarative Framework for Constrained Clustering

Dao, Thi-Bich-Hanh; Duong, Khanh-Chuong; Vrain, Christel

doi:10.1007/978-3-642-40994-3_27

Cited by 23 publications

(24 citation statements)

References 14 publications

Supporting

Mentioning

Contrasting

Unclassified

Order By: Relevance

“…This kind of relation can be realized by reified constraints, which were indeed used in our previous model [3]. However, a reified constraint is needed for each couple i < j, which implies that the number of reified constraints would be quadratic where D.lb is the lower bound, which initially can be the minimal dissimilarity between two points, and D.ub is the upper bound, which can be either +∞ or the value of D in the previous solution found.…”

Section: Diameter and Split Criteriamentioning

confidence: 98%

“…In a previous work [3], we have presented a CP model for this task. This model was based on a two-level representation: a set of variables for the assignment of a representative to each cluster and a set of variables for the assignment of a representative to each point.…”

Section: New Cp Model For Constrained Clusteringmentioning

confidence: 99%

“…When minimizing the diameter, in [3] we use heuristics provided by the algorithm FPF [6] to get a lower and an upper bound on the diameter without user constraints, and only an upper bound in the presence of user constraints. Such bounds allow to reduce the number of reified constraints that are put in the model.…”

Section: Optimization Criteriamentioning

confidence: 99%

“…3 An integer variable I c is introduced, its value is the index of the representative point of cluster c; the domain of I c is therefore the set of integers [1, n]. Let I be the array [I 1 , .…”

Section: Variablesmentioning

confidence: 99%

“…In this paper, we propose a new model for Constrained Clustering, still based on Constraint Programming, but significantly improved compared to the previous model [3]. In the previous model, two sets of variables were introduced, namely a variable for each cluster identifying a cluster by one of its points and a variable for each point expressing its assignment to a cluster.…”

mentioning

confidence: 98%

See 4 more Smart Citations

Constrained clustering by constraint programming

Dao

Duong

Vrain

2017

Artificial Intelligence

Self Cite

View full text Add to dashboard Cite

Constrained Clustering allows to make the clustering task more accurate by integrating user constraints, which can be instance-level or cluster-level constraints. Few works consider the integration of different kinds of constraints, they are usually based on declarative frameworks and they are often exact methods, which either enumerate all the solutions satisfying the user constraints, or find a global optimum when an optimization criterion is specified. In a previous work, we have proposed a model for Constrained Clustering based on a Constraint Programming framework. It is declarative, allowing a user to integrate user constraints and to choose an optimization criterion among several ones. In this article we present a new and substantially improved model for Constrained Clustering, still based on a Constraint Programming framework. It differs from our earlier model in the way partitions are represented by means of variables and constraints. It is also more flexible since the number of clusters does not need to be set beforehand; only a lower and an upper bound on the number of clusters have to be provided. In order to make the model-based approach more efficient, we propose new global optimization constraints with dedicated filtering algorithms. We show that such a framework can easily be embedded in a more general process and we illustrate this on the problem of finding the optimal Pareto front of a bi-criterion constrained clustering task. We compare our approach with existing exact approaches, based either on a branch-and-bound approach or on graph coloring on twelve datasets. Experiments show that the model outperforms exact approaches in most cases.

show abstract

Section: Diameter and Split Criteriamentioning

confidence: 98%

Section: New Cp Model For Constrained Clusteringmentioning

confidence: 99%

Section: Optimization Criteriamentioning

confidence: 99%

Section: Variablesmentioning

confidence: 99%

mentioning

confidence: 98%

See 3 more Smart Citations

Constrained clustering by constraint programming

Dao

Duong

Vrain

2017

Artificial Intelligence

Self Cite

View full text Add to dashboard Cite

show abstract

Constrained Clustering: Current and New Trends

Gançarski

Dao

Crémilleux

et al. 2020

A Guided Tour of Artificial Intelligence Research

Self Cite

View full text Add to dashboard Cite

Clustering is an unsupervised process which aims to discover regularities and underlying structures in data. Constrained clustering extends clustering in such a way that expert knowledge can be integrated through the use of user constraints. These guide the clustering process towards a more relevant result. Different means of integrating constraints into the clustering process exist. They consist of extending classical clustering algorithms, such as the well-known k-means algorithm; modelling the constrained clustering problem using a declarative framework; and finally, by directly integrating constraints into a collaborative process that involves several clustering algorithms. A common point of these approaches is that they require the user constraints to be given before the process begins. New trends in constrained clustering highlight the need for better interaction between the automatic process and expert supervision. This chapter is dedicated to constrained clustering. In particular, after a brief overview of constrained clustering and associated issues, it presents the three main approaches in the domain. It also discusses exploratory data mining by presenting models that develop interaction with the user in an incremental and collaborative way. Finally, moving beyond constraints, some aspects of user implicit preferences and their capture are introduced.

show abstract