Foreword to Fayyad, Piatetsky-Shapiro, Smyth, Uthurusamy (eds):
Keywords data mining, semantic models
Knowledge discovery is the most desirable end-product of computing. Finding new phenomena or enhancing our knowledge about them has a greater long-range value than optimizing production processes or inventories, and is second only to tasks that help preserve our world and our environment. It is not surprising that it is also one of the most difficult computing challenges to do well. Knowledge acquisition from experts often includes discovery as a byproduct, since the formalization often uncovers new linkages. Having a good model which lists and links all the candidate variables, including those that are not available in any databases helps in establishing sound knowledge acquisition schemes, although it cannot resolve problems of missing data. An important criterion is completeness of sets and subsets. When acquiring knowledge from data the problems of understanding capabilities and barriers must be dealt with so that the emerging knowledge discovery discipline can avoid the ups and downs that arise if great expectations are put forward, and then remain unfulfilled.