抄録
Special knowledge is needed in foundry shops. Especially, know-how on the diagnostics of casting defects is very important. Usually many companies have considerable documents on casting defects but it is still not clear now to use them effectively. The recent IT environment enables accumulation of such dicuments as electronic data naturally. In this study, a method to analyze and categorize document data on casting defects was developed using text-mining technology and knowledge map. Namely, delegate words for each casting characteristic are grouped and located on a two-dimensional map. Words related to the documents are highlighted on the map. This method enables visualization of the results of document classification, and has been found to be useful for classifying and retrieving documents to casting engineers.
