Using Support Vector Machine on EEG for Advertisement Impact Assessment

Area of Science:

• Neuroscience
• Marketing Science
• Human-Computer Interaction

Background:

• Current advertising impact assessment relies on indirect consumer behavior analysis or surveys, which are time-consuming and prone to bias.
• Existing methods often involve sunk costs and lack real-time feedback for campaign optimization.
• There is a need for rapid, quantitative, and unbiased methods to evaluate advertising effectiveness.

Purpose of the Study:

• To investigate a novel approach for assessing advertising impact using electroencephalography (EEG).
• To develop a quantitative and unbiased method for measuring the brain's response to advertisements.
• To provide a scalable methodology for advertising agencies to rapidly assess campaign effectiveness.

Main Methods:

• Utilized low-cost electroencephalography (EEG) headbands to record brain activity.
• Developed a Support Vector Machine (SVM) model to analyze EEG data and assess advertising impact.
• Conducted user experiments with 30 subjects exposed to 220 different advertisements.

Main Results:

• The proposed EEG-based method demonstrated desired performance in evaluating advertising impact.
• The SVM model successfully correlated brain activity with advertisement exposure.
• User experiments validated the feasibility of the approach.

Conclusions:

• The proposed EEG and SVM method offers a promising, rapid, and quantitative approach to assess advertising impact.
• This methodology can overcome the limitations of traditional assessment techniques, reducing bias and cost.
• Further development can lead to a general and scalable solution for the advertising industry.