Fast Subgroup Discovery for Continuous Target Concepts

Advanced Information Systems Engineering

Becker

Puppe

2013

Self Cite

Abstract. For the task of subgroup discovery, generalization-aware interesting measures that are based not only on the statistics of the patterns itself, but also on the statistics of their generalizations have recently been shown to be essential. A key technique to increase runtime performance of subgroup discovery algorithms is the application of optimistic estimates to limit the search space size. These are upper bounds for the interestingness that any specialization of the currently evaluated pattern may have. Until now these estimates are based on the anti-monotonicity of instances, which are covered by the current pattern. This neglects important properties of generalizations. Therefore, we present in this paper a new scheme of deriving optimistic estimates for generalization aware subgroup discovery, which is based on the instances by which patterns differ in comparison to their generalizations. We show, how this technique can be applied for the most popular interestingness measures for binary as well as for numeric target concepts. The novel bounds are incorporated in an efficient algorithm, which outperforms previous methods by up to an order of magnitude.

Section: Related Workmentioning

confidence: 99%

Difference-Based Estimates for Generalization-Aware Subgroup Discovery

Advanced Information Systems Engineering

Becker

Puppe

2013

Self Cite

“…-State-of-the-Art Algorithms: VIKAMINE comes with a variety of established and state-of-the-art algorithms for automatic subgroup discovery, e.g., , BSD [9], and SD-Map* [2]. A wide variety of popular interestingness measures can be used for binary, nominal, and numeric target concepts.…”

Section: Vikaminementioning

confidence: 99%

“…In this context, VIKAMINE provides state-of-the-art algorithmic implementations, cf. [2], for supporting the knowledge discovery and analysis, and enables an effective involvement of the domain experts.…”

Section: Exemplary Applicationsmentioning

confidence: 99%

“…In contrast to general purpose data mining systems, it is specialized for the task of subgroup discovery and pattern mining. It focuses on visual, interactive and knowledge-intensive methods and aims to integrate a distinctive set of features with an easy-to-use interface:-State-of-the-Art Algorithms: VIKAMINE comes with a variety of established and state-of-the-art algorithms for automatic subgroup discovery, e.g., , BSD [9], and SD-Map* [2]. A wide variety of popular interestingness measures can be used for binary, nominal, and numeric target concepts.…”

mentioning

confidence: 99%

See 1 more Smart Citation

VIKAMINE – Open-Source Subgroup Discovery, Pattern Mining, and Analytics

Atzmueller

Machine Learning and Knowledge Discovery in Databases

2012

Self Cite

Abstract. This paper presents an overview on the VIKAMINE 3 system for subgroup discovery, pattern mining and analytics. As of VIKAMINE version 2, it is implemented as rich-client platform (RCP) application, based on the Eclipse 4 framework. This provides for a highly-configurable environment, and allows modular extensions using plugins. We present the system, briefly discuss exemplary plugins, and provide a sketch of successful applications.Keywords: Pattern Mining, Subgroup Discovery, Analytics, Open-Source VIKAMINESubgroup discovery and pattern mining are important descriptive data mining tasks. They can be applied, for example, in order to obtain an overview on the relations in the data, for automatic hypotheses generation, and for a number of knowledge discovery applications. We present the VIKAMINE system for such applications.VIKAMINE is targeted at a broad range of users, from industrial practitioners to ML/KDD researchers, students, and users interested in knowledge discovery and data analysis in general. It features a variety of state-of-the-art automatic algorithms, visualizations, broad extensibility, and rich customization capabilities enabled by the Eclipse RCP environment. In contrast to general purpose data mining systems, it is specialized for the task of subgroup discovery and pattern mining. It focuses on visual, interactive and knowledge-intensive methods and aims to integrate a distinctive set of features with an easy-to-use interface:-State-of-the-Art Algorithms: VIKAMINE comes with a variety of established and state-of-the-art algorithms for automatic subgroup discovery, e.g., , BSD [9], and SD-Map* [2]. A wide variety of popular interestingness measures can be used for binary, nominal, and numeric target concepts.

“…Therefore, we propose a two step approach for tackling this problem: The first step uses pattern mining techniques, e.g., [1,2] to automatically generate a candidate set of potentially interesting descriptive tags. For this task, we present three different options for constructing target concepts.…”

Section: Introductionmentioning

confidence: 99%

Describing Locations Using Tags and Images: Explorative Pattern Mining in Social Media

Modeling and Mining Ubiquitous Social Media

Atzmueller

2012

Self Cite

Abstract. This paper presents an approach for explorative pattern mining in social media for describing image media based on tagging information and collaborative geo-reference annotations. We utilize pattern mining techniques for obtaining sets of tags that are specific for the specified point, landmark, or region of interest. Next, we show how these candidate patterns can be presented and visualized for interactive exploration using a combination of general pattern mining visualizations and views specialized on geo-referenced tagging data. We present a case study using publicly available data from the Flickr photo sharing platform.