An enhanced relevance criterion for more concise supervised pattern discovery

Großkreutz, Henrik; Paurat, Daniel; Rüping, Stefan

doi:10.1145/2339530.2339756

Cited by 9 publications

(5 citation statements)

References 26 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Another interesting question would be whether the approaches proposing closure operators on patterns taking the form of sets of sequences [17,18] could be combined with the concept of relevance. Finally, it would be interesting to investigate whether the notion of relevance can be further relaxed, following the ideas of [25,26].…”

Section: Discussionmentioning

confidence: 99%

A Relevance Criterion for Sequential Patterns

Großkreutz

Lang

Trabold

2013

Advanced Information Systems Engineering

View full text Add to dashboard Cite

The theory of relevance is an approach for redundancy avoidance in labeled itemset mining. In this paper, we adapt this theory to the setting of sequential patterns. While in the itemset setting it is suggestive to use the closed patterns as representatives for the relevant patterns, we argue that due to different properties of the space of sequential patterns, it is preferable to use the minimal generator sequences as representatives, instead of the closed sequences. Thereafter, we show that we can efficiently compute the relevant sequences via the minimal generators in the negatives. Unlike existing iterative or post-processing approaches for pattern subset selection, our approach thus results both in a reduction of the set of patterns and in a reduction of the search space -and hence in lower computational costs.

show abstract

Section: Discussionmentioning

confidence: 99%

A Relevance Criterion for Sequential Patterns

Großkreutz

Lang

Trabold

2013

Advanced Information Systems Engineering

View full text Add to dashboard Cite

show abstract

“…To address redundancy among the found subgroups, most previously proposed approaches encompass supervised pattern set mining (Bringmann and Zimmermann 2007), and methods based on relevance (Großkreutz et al 2012), and diversity (van Leeuwen and Knobbe 2011. Unlike diversity-based methods, the supervised pattern set mining objective is to find a fixed number of patterns, which must be chosen in At the time, relevance is limited to non-numeric targets.…”

Section: Redundancy Of Subgroup Sets and Subgroup Set Discoverymentioning

confidence: 99%

Robust subgroup discovery

Proença

Bäck

Leeuwen

2022

Data Min Knowl Disc

View full text Add to dashboard Cite

We introduce the problem of robust subgroup discovery, i.e., finding a set of interpretable descriptions of subsets that 1) stand out with respect to one or more target attributes, 2) are statistically robust, and 3) non-redundant. Many attempts have been made to mine either locally robust subgroups or to tackle the pattern explosion, but we are the first to address both challenges at the same time from a global modelling perspective. First, we formulate the broad model class of subgroup lists, i.e., ordered sets of subgroups, for univariate and multivariate targets that can consist of nominal or numeric variables, including traditional top-1 subgroup discovery in its definition. This novel model class allows us to formalise the problem of optimal robust subgroup discovery using the Minimum Description Length (MDL) principle, where we resort to optimal Normalised Maximum Likelihood and Bayesian encodings for nominal and numeric targets, respectively. Second, finding optimal subgroup lists is NP-hard. Therefore, we propose SSD++, a greedy heuristic that finds good subgroup lists and guarantees that the most significant subgroup found according to the MDL criterion is added in each iteration. In fact, the greedy gain is shown to be equivalent to a Bayesian one-sample proportion, multinomial, or t-test between the subgroup and dataset marginal target distributions plus a multiple hypothesis testing penalty. Furthermore, we empirically show on 54 datasets that SSD++ outperforms previous subgroup discovery methods in terms of quality, generalisation on unseen data, and subgroup list size.

show abstract

“…Such techniques can also be generalized for frequent patterns (cf., Refs ). For subgroup discovery, target‐closed representations can be formalized, cf., Ref for details. In that case, also an implicit redundancy management based on the subgroup descriptions is performed.…”

Section: Subgroup Set Selectionmentioning

confidence: 99%

“…In addition, such a method can also be implemented in an optional post‐processing step. Furthermore, Großkreutz et al introduces delta‐relevance which provides a more relaxed definition of coverage (essentially trading off precision versus simplicity) with the overall goal of summarizing relevant patterns even more.…”

Section: Subgroup Set Selectionmentioning

confidence: 99%

Subgroup discovery

Atzmueller

2015

WIREs Data Min & Knowl

185

102

View full text Add to dashboard Cite

Subgroup discovery is a broadly applicable descriptive data mining technique for identifying interesting subgroups according to some property of interest. This article summarizes fundamentals of subgroup discovery, before that it also reviews algorithms and further advanced methodological issues. In addition, we briefly discuss tools and applications of subgroup discovery approaches. In that context, we also discuss experiences and lessons learned and outline some of the future directions in order to show the advantages and benefits of subgroup discovery.

show abstract

An enhanced relevance criterion for more concise supervised pattern discovery

Cited by 9 publications

References 26 publications

A Relevance Criterion for Sequential Patterns

A Relevance Criterion for Sequential Patterns

Robust subgroup discovery

Subgroup discovery

Contact Info

Product

Resources

About