EEG Data Analysis with Artificial Intelligence for Energy Data Visualisations

Energy efficiency becomes crucial because it saves money, reduces consumption, and lowers demand. Smart home technology holds tremendous potential, but the key to success lies in user interaction. The usability of smart home data visualisations is inadequate suggesting limited user engagement. To engage users with energy-saving technologies, accurate real-time data and effective presentation methods are required. Traditional methods of gathering user preferences via surveys and questionnaires are insufficient for comprehending energy user preferences and displaying energy consumption statistics. Brain-Computer Interface (BCI) combined with Artificial Intelligence (AI) can assist in analysing people's preferences for external stimuli, reducing challenges in interacting with energy data, and improving the usability of smart home systems. Because of the distinctive stimuli, Electroencephalography (EEG) data collection for energy data visualisations as part of BCI systems can provide a unique contribution to emotion recognition from brain signals and emotion classification for these stimuli using AI algorithms. This thesis focuses on acquiring EEG data for energy data visualisations by collecting subjective responses related to emotions using Self-Assessment Manikins (SAMs). Additionally, Generative Adversarial Networks (GANs) are leveraged to boost the quantity of data and enhance performance. The GAN model samples are combined with empirical EEG data and classified using AI algorithms to determine energy data visualisation preferences based on the valence and arousal levels of participants. Furthermore, channel effectiveness in assessing user responses to data visualisations is evaluated using Two-Dimensional (2-D) representations and Deep Learning (DL). The results show a valence classification accuracy of 99.53% in EEG classification and a maximum valence accuracy of 100% when using DL features and the Spectrogram method with the F8 channel in EEG channel analysis.