One may ask what kind of uses are there for data mining techniques in industrial automation. There are many, some of them are obvious and directly related to that application area, some are less obvious and they may have common ground with many other areas. One potential application could be based on the Self-Organizing Map slightly modifying the basic idea of the WEBSOM method. In the WEBSOM, contexts for words are analyzed, and that information is used in encoding the documents. In an industrial setting one might associate processes and their states with documentation. The process and state descriptions might consist both of continuous parameters values and discrete classifications. For each document in, say, the collection of manuals, the description vectors would then be used as the contexts for the documents. This would lead into an organized map of the manuals that could be browsed, or that could be used automatically; the process parameters indicate the best-matching node in the map, thus, the document that best describes the process state would be readily available at any moment.
The scope of the potential application is naturally wide while industrial automation systems often produce vast amount of detailed information from which generalizations would be potentially very useful. One may, for example, find new solutions for more efficient use of energy or raw materials, or ways to reduce the amount of waste materials, based on the refined information.