An Unusual Approach to Basic Challenges of Data Mining
Thomas H. Lenhard, Michal Greguš · 2015
In this paper, we will give a short overview about the complexity and the challenges of basic steps when building a Data Mining System. Such a work cannot be realized without empirical experience. For this reason the essence of many projects has found its way into this paper. After the introduction that explains premises of Data Mining that are often underrated or ignored, it is said, that one of the biggest challenges in Data Mining is identification of data inside data sources. The paper shows several kinds of data sources that may include internal or external data and that may be of very different types of data bases or data files. It also explains kinds of traps and difficult challenges of the first steps of Data Mining. Several constellations and situations will be discussed and a method, which is described in literature, will be identified as being out of date. Finally a method for identifying tables, attributes and constraints inside a relational data base is presented, that is effective and highly efficient: Sniffing!