A Relational View of Pattern Discovery

Giacometti, Arnaud; Marcel, Patrick; Soulet, Arnaud

doi:10.1007/978-3-642-20149-3_13

Cited by 5 publications

(5 citation statements)

References 23 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Although Table 1 is an extremely small sample of measures, we observe that the evaluation complexity of pattern mining methods has increased over the past decades. Interestingly, we also note that the evaluation complexity of global constraints [6,13] (or advanced measures [26]) is greater than those of local constraints (or absolute measures).…”

Section: Evaluation Complexity Of Interestingness Selectorsmentioning

confidence: 74%

See 1 more Smart Citation

How Your Supporters and Opponents Define Your Interestingness

Crémilleux

Giacometti

Soulet

2019

Machine Learning and Knowledge Discovery in Databases

Self Cite

View full text Add to dashboard Cite

How can one determine whether a data mining method extracts interesting patterns? The paper deals with this core question in the context of unsupervised problems with binary data. We formalize the quality of a data mining method by identifying patterns-the supporters and opponents-which are related to a pattern extracted by a method. We define a typology offering a global picture of the methods based on two complementary criteria to evaluate and interpret their interests. The quality of a data mining method is quantified via an evaluation complexity analysis based on the number of supporters and opponents of a pattern extracted by the method. We provide an experimental study on the evaluation of the quality of the methods.

show abstract

Section: Evaluation Complexity Of Interestingness Selectorsmentioning

confidence: 74%

“…The concept of global constraints has been informally defined in [6]. This notion has been formalized in [13] by defining a relational algebra extended to pattern discovery. Our framework provides a broader and more precise formal definition, especially to better analyze the interrelationships between the patterns.…”

Section: Related Workmentioning

confidence: 99%

How Your Supporters and Opponents Define Your Interestingness

Crémilleux

Giacometti

Soulet

2019

Machine Learning and Knowledge Discovery in Databases

Self Cite

View full text Add to dashboard Cite

show abstract

“…Techniques for mining itemsets and n-ary patterns (i.e., the combination of n patterns) have been proposed. Looking for declarative techniques in pattern mining was also recently investigated in [11]: by using relational algebra, the authors propose an algebraic framework for pattern discovery for expressing a wide range of queries.…”

Section: Related Workmentioning

confidence: 99%

A constraint language for declarative pattern discovery

Métivier

Boizumault

Crémilleux

et al. 2012

Proceedings of the 27th Annual ACM Symposium on Applied Computing

View full text Add to dashboard Cite

Discovering pattern sets or global patterns is an attractive issue from the pattern mining community in order to provide useful information. By combining local patterns satisfying a joint meaning, this approach produces patterns of higher level and thus more useful for the end-user than the usual local patterns. In parallel, recent works investigating relationships between data mining and constraint programming (CP) show that the CP paradigm is a powerful framework to model and mine patterns in a declarative and generic way. We present a constraint-based language which enables us to define queries in a declarative way addressing patterns sets and global patterns. By specifying what the task is, rather than providing how the solution should be computed, it is easy to process by stepwise refinements to successfully discover global patterns. The usefulness of the approach is highlighted by several examples coming from the clustering based on associations. All primitive constraints of the language are modeled and solved using the SAT framework. We illustrate the efficiency of our approach through several experiments.

show abstract

“…Even if the data type is a complex data type, it is always possible to un-nest it into a relational-like structure -and, in any case, it is important that what matters here, i.e., the model components, are explicitly modeled as attributes. Attempts to relationaly code mining results already exist [68].…”

Section: Models Model Components and Data-to-model Mappingsmentioning

confidence: 99%

Beyond roll-up’s and drill-down’s: An intentional analytics model to reinvent OLAP

Vassiliadis

Marcel

Rizzi

2019

Information Systems

Self Cite

View full text Add to dashboard Cite

This paper structures a novel vision for OLAP by fundamentally redefining several of the pillars on which OLAP has been based for the last 20 years. We redefine OLAP queries, in order to move to higher degrees of abstraction from roll-up's and drill-down's, and we propose a set of novel intentional OLAP operators, namely, describe, assess, explain, predict, and suggest, which express the user's need for results. We fundamentally redefine what a query answer is, and escape from the constraint that the answer is a set of tuples; on the contrary, we complement the set of tuples with models (typically, but not exclusively, results of data mining algorithms over the involved data) that concisely represent the internal structure or correlations of the data. Due to the diverse nature of the involved models, we come up (for the first time ever, to the best of our knowledge) with a unifying framework for them, that places its pillars on the extension of each data cell of a cube with information about the models that pertain to it -practically converting the small parts that build up the models to data that annotate each cell. We exploit this data-to-model mapping to provide highlights of the data, by isolating data and models that maximize the delivery of new information to the user. We introduce a novel method for assessing the surprise that a new query result brings to the user, with respect to the information contained in previous results the user has seen via a new interestingness measure. The individual parts of our proposal are integrated in a new data model for OLAP, which we call the Intentional Analytics Model. We complement our contribution with a list of significant open problems for the community to address.

show abstract

A Relational View of Pattern Discovery

Cited by 5 publications

References 23 publications

How Your Supporters and Opponents Define Your Interestingness

How Your Supporters and Opponents Define Your Interestingness

A constraint language for declarative pattern discovery

Beyond roll-up’s and drill-down’s: An intentional analytics model to reinvent OLAP

Contact Info

Product

Resources

About