An Intelligent EEG-based Advertisement Recommendation System using Correlative Capsule Networks and Similarity Matching

Abstract

This paper describes the design and implementation of a new Brainwave Driven Advertisement Recommendation System using state of the art Emotive Insight hardware, Brain Computer Interface (BCI) and Artificial Intelligence (AI). The technology is primarily aimed at improving the precision and personalization of ad recommendations by directly harnessing people’s neural responses. The Emotive Insight headset can record brainwave frequencies including alpha, beta, theta, delta and gamma in real time. This rich data set is carefully preserved in both edf formats and is the basis for a thorough analysis. The participants saw three separate advertisements: Snickers, Dairy Milk and 5-star. To test the effect of each advertisement, they measure their brain responses simultaneously. Then the users comment on what ads they like best, tying subjective beliefs to brain wave patterns. Data Analysis: The data analysis stage involves meticulous preprocessing, artifact removal, and feature extraction to guaranty the extraction of meaningful insights. We build a strong predictive model using various machine learning approaches. The model is evaluated by the accuracy, precision, recall and F1 score metrics.