ABSTRACT
The practical purpose of this entry is to characterize the information to be collected and organized during the Data Evaluation Step of a data mining spiral. Rather than developing these ideas in prose from which the reader must extract actionable chunks, the material is presented as a topically organized checklist of questions to be addressed. This entry also describes some techniques for conducting the preliminary analysis of domain data; that is, analysis which is performed before the Feature Extraction Step. The aspect that distinguishes this genre from others is that here we are searching data for instances of unknown patterns; patterns that are not well characterized, and for which representative examples are not available.
