Some Industrial Applications of Text Mining

Abstract

Three industrial applications of text mining will be presented requiring different methodologies. The first application used a classification approach in order filter documents relevant for personal profiles from an underlying document collection. The second application combines cluster analysis with statistical trend analysis in order detect emerging issues in manufacturing. In the third application a combination of static term indexing and dynamic singular value computation is used to drive similarity search in a large document collection. All of these approaches require a knowledgeable human to be part of the process, the goal is not an automatic knowledge understanding but using text mining technology in order to enhance the productivity of existing business processes.