INVESTIGATING GENDER VARIATIONS IN DREAM EMOTION PREDICTION USING EEG SIGNAL AND RANDOM FOREST ALGORITHM

Abstract

Dreams offer a unique insight into the subconscious emotional and psychological states of individuals. While EEG-based emotion recognition has been explored in various contexts, the gender differences in dream emotion classification remain largely unexplored. This study investigates the prediction of dream emotions (Positive, Negative, Neutral) from EEG signals using the Random Forest algorithm, with a focus on gender-based distinctions [1]. EEG signals were collected from male and female participants during REM sleep, and a range of statistical and frequency-domain features were extracted from standard EEG channels. The Random Forest classifier, trained on these features, demonstrated high accuracy in classifying dream emotions, with slightly better performance for females [2]. Gender-wise analysis revealed that males exhibited higher classification accuracy for positive emotions, while females showed stronger recognition of negative and neutral emotions. These findings underscore the importance of considering gender-specific factors in dream emotion research and pave the way for future advancements in personalized emotion recognition systems [3]. Future research will explore larger, more diverse datasets, advanced deep learning methods, and multimodal signal integration to further improve the model's generalizability and predictive accuracy.