A Framework for Semi-Supervised Learning Based on Subjective and Objective Clustering Criteria

Halkidi, Maria; Gunopulos, Dimitrios; Kumar, N. V. Narendra; Vazirgiannis, M.; Domeniconi, Carlotta

doi:10.1109/icdm.2005.4

Cited by 13 publications

(11 citation statements)

References 8 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…A further method of the same category appears in Halkidi et al (2005): constraint violation penalties are incorporated to the distance metric and then combined with a modified objective function. The authors use a hill-climbing algorithm to find an optimal solution.…”

Section: Hybrid Methodsmentioning

confidence: 99%

“…5, we disclosed some labels, identified queries that had the same label and combined them into Must-Link constraints. This approach was also used in Wagstaff and Cardie (2000), Halkidi et al (2005), Ruiz et al (2007a). For each QS i , we generated a set of 200 Must-Link constraints ML i, [200] Fig.…”

Section: Methodsmentioning

confidence: 99%

“…Similarly to Wagstaff et al (2001) and Halkidi et al (2005), we use the class labels of the test datasets to generate constraints. In particular, we select pairs of data points randomly and check the label of each pair: if the two data points have the same label, we generate a Must-Link constraint for them, otherwise we generate a Cannot-Link constraint.…”

Section: Generating Constraintsmentioning

confidence: 99%

“…We then turn to algorithms that use the constraints to "learn" a new distance metric for the clustering algorithm (Basu et al 2004b;Bilenko et al 2004;Halkidi et al 2005). Under this approach, which is often termed distance-based, clustering is done in a conventional way, using the new distance function.…”

Section: Different Ways Of Instance-level Constraint Enforcementmentioning

confidence: 99%

“…Therefore, most constraint-based methods calculate the average performance achievements of the results using random sets of constraints (Wagstaff and Cardie 2000;Wagstaff et al 2001;Halkidi et al 2005;Ruiz et al 2007a).…”

Section: Determining the Set Of Constraintsmentioning

confidence: 99%

See 4 more Smart Citations

Density-based semi-supervised clustering

Ruiz

Spiliopoulou

Menasalvas

2009

Data Min Knowl Disc

View full text Add to dashboard Cite

Semi-supervised clustering methods guide the data partitioning and grouping process by exploiting background knowledge, among else in the form of constraints. In this study, we propose a semi-supervised density-based clustering method. Density-based algorithms are traditionally used in applications, where the anticipated groups are expected to assume non-spherical shapes and/or differ in cardinality or density. Many such applications, among else those on GIS, lend themselves to constraint-based clustering, because there is a priori knowledge on the group membership of some records. In fact, constraints might be the only way to prevent the formation of clusters that do not conform to the applications' semantics. For example, geographical objects, e.g. houses, separated by a borderline or a river may not be assigned to the same cluster, independently of their physical proximity. We first provide an overview of constraint-based clustering for different families of clustering algorithms. Then, we concentrate on the density-based algorithms' family and select the algorithm DBSCAN, which we enhance with Must-Link and Cannot-Link constraints. Our enhancement is seamless: we allow DBSCAN to build temporary clusters, which Responsible editor: Eamonn Keogh. 123 346 C. Ruiz et al.we then split or merge according to the constraints. Our experiments on synthetic and real datasets show that our approach improves the performance of the algorithm.

show abstract

Section: Hybrid Methodsmentioning

confidence: 99%

Section: Methodsmentioning

confidence: 99%

Section: Generating Constraintsmentioning

confidence: 99%

Section: Different Ways Of Instance-level Constraint Enforcementmentioning

confidence: 99%

Section: Determining the Set Of Constraintsmentioning

confidence: 99%

See 3 more Smart Citations