A framework for redescription set construction

Abstract

Redescription mining is a field of knowledge discovery that aims at finding different descriptions of similar subsets of instances in the data. These descriptions are represented as rules inferred from one or more disjoint sets of attributes, called views. As such, they support knowledge discovery process and help domain experts in formulating new hypotheses or constructing new knowledge bases and decision support systems. In contrast to previous approaches that typically create one smaller set of redescriptions satisfying a pre-defined set of constraints, we introduce a framework that creates large and heterogeneous redescription set from which user/expert can extract compact sets of differing properties, according to its own preferences. Construction of large and heterogeneous redescription set relies on CLUS-RM algorithm and a novel, conjunctive refinement procedure that facilitates generation of larger and more accurate redescription sets. The work also introduces the variability of redescription accuracy when missing values are present in the data, which significantly extends applicability of the method. Crucial part of the framework is the redescription set extraction based on heuristic multi-objective optimization procedure that allows user to define importance levels towards one or more redescription quality criteria. We provide both theoretical and empirical comparison of the novel framework against current state of the art redescription mining algorithms and show that it represents more efficient and versatile approach for mining redescriptions from data.

Targeted and contextual redescription set exploration

“…In future work, we intend to extend the applicability of the tool for exploration of other kinds of rule sets, such as association rules and subgroups. The tool will also be extended to work with different redescription set construction techniques (Kalofolias et al 2016;Mihelčić et al 2017a), and enable the use of statistical tests within different aspects to facilitate the discovery process. One potentially interesting direction of future research is to explore the potential of developed algorithms and adapted visualizations in an interactive redescription mining setting.…”

“…Our approach provides a framework that allows incorporating ideas from various pattern set mining and pattern selection approaches, designed to create smaller, compressed or userpreferred set of patterns, into redescription set mining and exploration process. Pattern set mining approaches, such as Xin et al 2005 2017can be used to construct sets of new redescriptions describing the selected subsets of entities/attributes of interest, whereas pattern selection methods (Xin et al 2005;Knobbe and Ho 2006;Pei et al 2007;Ouali et al 2017;Kalofolias et al 2016;Mihelčić et al 2017a) can be used to perform automated and fine-grained selection of redescriptions from a selected subset of redescriptions. Techniques for performing redescription mining with entity/attribute constraints already exist (Zaki and Ramakrishnan 2005;Mihelčić et al 2017b).…”

“…Choice of methodology depends on user preferences and the quality of final set of redescriptions depends on predefined redescription and redescription set quality measures. One directly applicable, scalable methodology that can be used for construction of smaller sets of redescriptions, using user-selected subset of redescriptions and allowing definition of user-provided preference information is the generalized redescription set construction procedure (Mihelčić et al 2017a).…”

Using redescription mining to relate clinical and biological characteristics of cognitively impaired and Alzheimer’s disease patients

“…Choice of redescription accuracy measure. Since the data contains missing values, we used the query non-missing Jaccard index, introduced in [30], and further explained in [45] to evaluate redescriptions. The query non missing Jaccard index is defined as:…”

Approaches for Multi-View Redescription Mining

“…As a part of the related work, we present a slightly modified generalized version of the CLUS-RM algorithm [27] which we call the GCLUS-RM. The generalised CLUS-RM algorithm (GCLUS-RM), presented in Algorithm 1, contains memory constraints on the maximal size of the redescription set and allows using an arbitrary, rule-transformable, model M obtained using some multi-target regression (multi-label classification), machine learning algorithm Alg (lines 1 to 10).…”