A novel subject-independent deep learning approach for user behavior prediction in electronic markets based on electroencephalographic data

Based on the work by Buettner (2017) showing a personality-based recommender system for electronic markets using social media data, we extend the work by proposing a novel deep learning-based engine to predict the user’s personality just based on electroencephalographic brain data. As brain-computer interfaces and hybrid intelligence devices enable access to human brains, using electroencephalographic brain data becomes more relevant in future. Contrary to the majority view of previous research, our results show that there is a link between personality traits and brain features of a user. With a four times higher probability of correctly predicting the personality of an independent user compared to naive prediction, we demonstrate the possibility of predicting a user’s personality based on their brain information and thus showing a new reliable approach for marketing purposes in electronic markets.