The current report seeks to understand the selection dynamics of patent classes (CPCs) in Europe by employing the methodology developed by Acemoglu et al. (2016) to predict future patenting. Their research focuses on citation networks measuring the knowledge flows across technologies and uses theses to estimate future volumes of patents per CPC during 1995-2004 in the United States. In our current analysis we replicate their results using the European patent database for the years 2005-2014, and likewise demonstrate that the innovation networks have significant predictive power over future patenting in Europe. Furthermore, we improve their methodology by accounting for more complex interactions between CPCs. Finally, we discuss their implications for developing a selection-dynamics model grounded in evolutionary theory.