Text mining for corporate information retrieval

Authoring technical documentation of products is a time-consuming process for industry. Furthermore,products are typically not designed in isolation, but belong to product lines, in which different products sharefeatures and operating instructions. Therefore, authoring technical documentation should be enabled to buildon existing documentation components that are retrieved and adapted to different, but similar, products.Innovatia is an industry leader in supporting companies re-use and deploy in electronic form existingdocumentation materials. The proposed project will continue to investigate the application of text mining andtext visualization techniques to the problem of management and re-use of technical documentation.Different text similarity methods will be applied, aiming to improve the ability of Innovatia's authoring systemto identify and cluster together similar document components, even if they use different wording to expresssimilar concepts. In the proposed phase of the project, authors from Innovatia will use the implementedextensions to the company's Content Miner system to evaluate their effectiveness in the authoring task.Additional problems that will be addressed in this phase include the automatic evaluation of languagenon-uniformity, and the application of visual text analytics techniques to deriving insight from help-desktickets.