| Abstrakt: |
Emotion recognition utilizing MindWave signals and neural networks presents a substantial challenge due to the inherent complexity of human emotions and the variability of individual brainwaves. The selection of the appropriate algorithm, dictated by the problem and available data, necessitates an understanding of each algorithm's unique strengths and weaknesses. Previous studies have predominantly focused on the classification of emotions through EEG signals employing various standalone neural network algorithms. However, our study fills a notable research gap by integrating Convolutional Neural Networks (CNN) and Gated Recurrent Units (GRU). This innovative combination yields improved testing performance and accuracy, setting a benchmark in the realm of emotion recognition. The process encompasses the collection of MindWave data, the elimination of noise through preprocessing, the extraction of features indicative of emotional states, and the training of a neural network using labeled data. Finally, the network's accuracy is evaluated on novel data. By addressing the unique challenges and complexities associated with emotion classification using EEG signals, this study provides a promising and advanced approach towards the understanding and recognition of human emotions, paving the way for potential real-world applications. [ABSTRACT FROM AUTHOR] |
