Conceptual Knowledge Discovery in DatabasesUsing Formal Concept Analysis

In this paper we discuss Conceptual Knowledge Discovery in Databases (CKDD) as it is developing in the eld of Conceptual Knowledge Processing (cf. 29],,30]). Conceptual Knowledge Processing is based on the mathematical theory of Formal Concept Analysis which has become a successful theory for data analysis during the last 18 years. This approach relies on the pragmatic philosophy of Ch.S. Peirce 15] who claims that we can only analyze and argue within restricted contexts where we always rely on pre-knowledge and common sense. The development of Formal Concept Analysis led to the software system TOSCANA, which is presented as a CKDD tool in this paper. TOSCANA is a exible navigation tool that allows dynamic browsing through and zooming into the data. It supports the exploration of large databases by visualizing conceptual aspects inherent to the data. We want to clarify that CKDD can be understood as a human-centered approach of Knowledge Discovery in Databases. The actual discussion about human-centered Knowledge Discovery is therefore brieey summarized in Section 1. 1 Human-Centered Knowledge Discovery Knowledge Discovery in Databases (KDD) is aimed at the development of methods , techniques, and tools that support human analysts in the overall process of discovering useful information and knowledge in databases. Many real-world knowledge discovery tasks are both too complex to be accessible by simply applying a single learning or data mining algorithm and too knowledge-intensive to be performed without repeated participation of the domain expert. Therefore , knowledge discovery in databases is considered an interactive and iterative process between a human and a database that may strongly involve background knowledge of the analyzing domain expert. This process-centered view of KDD is the overall theme and contribution of the volume\Advances in Knowledge Discovery and Data Mining" 7]. According to R.S. Brachman and T. Anand 3], much attention and eeort has been focused on the development of data-mining techniques but only a minor eeort has been devoted to the development of tools that support the analyst in the overall discovery task. They see a clear need to emphasize the processori-entation of KDD tasks and argue in favor of a more human-centered approach