Demographic Market Segmentation on Short Banking Movement Descriptions Applying Natural Language Processing

Banking movement descriptions can be a valuable type of short texts for knowledge extraction with application in finance and social studies. Conventional research on text mining has mostly been applied to medium-sized documents. Knowledge extraction from banking movement descriptions is challenging due to the lack of meaningful textual data and their ad-hoc terminology. In this work we present a clustering analysis on short banking movement descriptions based on Natural Language Processing techniques. We exploit the knowledge in an experimental data set composed of almost 20,000 real banking transactions that have been anonymised as required by European data protection regulations. At the end, we were able to extract five distinctive user clusters with similar demographics. Our approach has potential applications in Personal Finance Management.