Vyhledávání znalostí z dat a statistika

Abstract: Knowledge Discovery in Databases and Statistics This article explains the relationship among statistics, data mining (DM) and knowledge discovery in databases (KDD). KDD and DM became important tools to the solution of the problem of knowledge discovery from data. In the frame of the formalized process of knowledge extraction there is more important for the user to stress his attention to what kind of data is used than attention to which kind of methods is used for extraction. A competent user of the results extracted with the KDD system plays the key role. The selection of valid data, preprocessing of data, the use of the tools that are able to detect violation of assumptions, and the tools that allow visualization of the results should be implemented in the KDD system. Besides the benefits of the KDD process, also some questionable aspects of KDD implementation are mentioned here: for example social impacts resulting from the hidden collection of the clickstreams information.