A first step towards automatic identification and provision of user-specific knowledge: A verification of the feasibility of automatic text classification using the example of standards

Abstract

The product development process faces major challenges. One major challenge is the time-efficient retrieval of information from engineering documents. In times of machine learning - as an enabler as well as a driver of digitization - new possibilities are opening up to meet the need for intelligent provision of knowledge and to make the flood of information in the product development process more manageable. To achieve the superordinate – future – goal of automatic and user-specific identification and provision of knowledge, a step-by-step approach is required. Thus, existing approaches dealing with the solution of classification problems by ontology-based methods are evaluated, transferred to the domain of standards, and extended. As an initial step, the technical feasibility of automatic text classification for the assignment of business departments to corresponding contents of documents will be examined within the scope of this research work, so that users are provided with the possibility of support when searching for information. Based on this, future research activities will primarily deal with the development of a role model so that, with the help of defined criteria, an unambiguous assignment can be made between content from different engineering documents and a "person of interest".