Hier finden Sie wissenschaftliche Publikationen aus den Fraunhofer-Instituten.

Identification of interdisciplinary ideas

: Thorleuchter, Dirk; Poel, Dirk van den


Information processing and management 52 (2016), No.6, pp.1074-1085
ISSN: 0306-4573
Journal Article
Fraunhofer INT ()
text mining; Latent Semantic Indexing; idea mining; clustering; classification

Literature shows interdisciplinary research as an essential driver for innovation. Ideas that are used as a starting point for this research are of an interdisciplinary nature because they combine aspects from different disciplines. The identification of interdisciplinary ideas at an early stage enables the start of interdisciplinary research and thus, it enables advances to be made in the innovation process. We propose a new methodology that combines semantic clustering and classification to estimate the interdisciplinary nature of ideas from a set of given ideas. The set is created automatically by use of an existing idea mining approach. Ideas from this set are semantically clustered to obtain concepts that are latent in the data. The relationship between each concept and each discipline pair from a set of given disciplines is calculated. Based on the degree of relationship, concepts are used to represent the interdisciplinary field spanned by the two disciplines. The ideas standing behind these concepts are identified as interdisciplinary ideas. As a result, the proposed methodology enables an estimation of the interdisciplinary nature of given ideas. The results might be helpful for researchers as well as for decision makers in the field of innovation management.