Abstract:

Finding valuable conclusions in all sorts of data is the task of the 21th century – creating the tools to do it automatically the current challenge. Especially large-scale research clusters are confronted with the challenge of managing their research efforts. These scientific networks usually consist of different disciplines with their own research projects. Leaving the question open, of how those individual research projects have developed over time. This historical examination could be of great benefit, for example, for future management, organization and the further strategic orientation of the research cluster. To approach this goal, an automated analysis of those large amounts of papers that are continuously produced in a research networks, offer potential for a time-based evaluation using data analysis. Past research projects have successfully shown, that text-mining algorithms offer one way to display and improve research collaboration in those research networks.

The following paper describes a technical concept, using a statistical approach based on text-mining algorithms with a combination of a time series model to display the evolution of those research topics in a research network. First, Text mining algorithms are used to process those research papers of a research network. The necessary pre-processing steps are described and leading to the generation of topic models. These topic models are used to have a textual identification of those research projects. Secondly, the identified research topics are put in relation to their publication date in order to be able to represent these research topics and their development over time. Time base models will hereby be considered as a feasible method. Finally, we will discuss possible forms of visualization that are adapted to the problem and can improve further cooperation in the research network. The document corpus of the research network Cluster of Excellence "Integrative Production Technology for High-Wage Countries" will be used as a data basis.

Keywords:

Text Mining, Research Progress, Data Analytics, Production Engineering.