Electroencephalography based Like/Dislike Classification Using Brain Network Features for Applications in Neuromarketing

Neuromarketing is an emerging interdisciplinary area that combines the two disciplines of marketing and neuroscience. Brain networks are graph-theoretic knowledge commonly used in the medical field. After the data is preprocessed, we set the electrode points as nodes and use the Pearson coefficient to construct the relationship between nodes, and extract three commonly used brain network attributes: node degree, clustering coefficient and global efficiency. A consumer like/dislike prediction system is constructed based on brain network features, and its classification accuracy reaches 84%. The results in this study indicated that brain networks could be used to distinguish like and dislike with potential applications in neuromarketing.