Types and forms of data

Knowledge discovery methods analyse data of quite diverse types such as cross sections, time series, texts, or multimedia data. We first summarize dimensions being important for identifying subtypes of these main general data types that require special analysis approaches. Thus specifically for complex, i.e. large, high dimensional, multirelational, or dynamic datasets, problems occur with current views on data developed in statistics and other disciplines. Next, we deal with some organization forms. Data are not collected as an unregulated set of numbers or strings, but appear in some organized form. Starting at the classical form used in data analysis for organizing cross sectional data, the flat rectangular organization, we proceed to more complex forms and deal with time and space referenced data. Finally, we treat data aggregation levels and meta data. A classification of data types is useful to select appropriate mining tasks, analysis methods, and data management solutions for an application.