Benchmarking of Scientific Research Clusters by Use of Text Mining Algorithms on Textual Artefacts

摘要

The development of text mining algorithms is far reaching and thus, new application areas arise, inter alia in the framework of applying modern information technologies (e.g. big data analytics) in order to answer social and economic (research) questions (digital humanities). In this paper existing algorithms to reveal results in the context of benchmarking scientific research clusters are combined. For conducting the analysis, two data corpora are gathered. The first data corpus contains abstracts of publications according to the used application case of the Cluster of Excellence Tailor-Made Fuels from Biomass 'TMFB' at RWTH Aachen University. The second data corpus -- based on the analysis of the first data corpus -- contains content-related abstracts of research belonging to the relevant scientific community. In the framework of processing the gathered data, information extraction, information retrieval and named entity recognition among others are utilized for analyzing the publications in order to derive useful findings. Furthermore association mining findings are introduced. As an excerpt of the results, the identification of thematic alignments and developments over time are presented for both corpora and are subsequently compared.