Supporting business process improvement with natural language processing: A model-based approach

The research area of business process improvement (BPI) focuses on the development of methods and techniques for redesigning processes as well as improving the process performance accordingly. Thereby one particular challenge lies in the classi®cation of customer feedback that is received via various online and of ̄ine channels. In the paper at hand we present an approach how customer feedback that is issued via social media can be classi®ed using natural language processing. For this purpose, a domain-speci®c modeling method has been designed that permits to specify the processing steps in visual form and that can integrate the results in existing modeling methods for BPI. For a ®rst evaluation, the approach has been prototypically implemented on the ADOxx-based SeMFIS platform and assessed using a SWOT analysis.