Visual data analysis with computational intelligence methods

Bulletin of the Polish Academy of Sciences. Technical Sciences

R. KruseDepartment of Knowledge and Language Engineering, Faculty of Computer Science Otto-von-Guericke-University of Magdeburg, D-39016 Magdeburg, GermanyOther articles by this author:De Gruyter OnlineGoogle Scholar¹  M. SteinbrecherDepartment of Knowledge and Language Engineering, Faculty of Computer Science Otto-von-Guericke-University of Magdeburg, D-39016 Magdeburg, GermanyOther articles by this author:De Gruyter OnlineGoogle Scholar¹ 

[1] Department of Knowledge and Language Engineering, Faculty of Computer Science Otto-von-Guericke-University of Magdeburg, D-39016 Magdeburg, Germany

关键词: Keywords: visual data analysis;    computational intelligence methods;   

DOI  :  10.2478/v10175-010-0037-z

学科分类：工程和技术（综合）

来源: Polska Akademia Nauk * Centrum Upowszechniania Nauki / Polish Academy of Sciences, Center for the Advancement of Science

【 摘 要 】

Visual data analysis is an appealing and increasing field of application. We present two related visual analysis approaches that allow for the visualization of graphical model parameters and time-dependent association rules. When the graphical model is defined over purely nominal attributes, its local structure can be interpreted as an association rule. Such association rules comprise one of the most prominent and wide-spread analysis techniques for pattern detection, however, there are only few visualization methods. We introduce an alternative visual representation that also incorporates time since patterns are likely to change over time when the underlying data was collected from real-world processes. We apply the technique to both an artificial and a complex real-life dataset and show that the combined automatic and visual approach gives more and faster insight into the data than a fully-automatic approach only. Thus, our proposed method is capable of reducing considerably the analysis time.

【 授权许可】

Unknown