Is consideration of background knowledge in Data Driven Solutions possible at all?

Abstract

In Data Driven Solutions using CBR or Data Mining approaches the optimal results can be achieved if one can consider and use, besides available data and texts, all other available information sources like general and background knowledge. Formalization and integration of such kind of knowledge in the knowledge extracted from data and texts is not, however, a simple task. For this reason, a lot of approaches, among them Bayesian Networks and Inductive Logic Programming, have been suggested in the literature to solve this problem. In the talk, this topic is discussed pragmatically by reviewing the personal experiences of the speaker in the last 20 years using concrete examples from the automotive industry. © Springer-Verlag Berlin Heidelberg 2006.