Abstract
The traditional music recommendation algorithm is an algorithm that measures the similarity of user preferences, or an algorithm that distinguishes types according to music styles and genres, which cannot meet the needs of emotional induction for different personality types. It is necessary to study the emotional induction of varying personality types by Western music based on the Internet of Things (IoT) technology. Electroencephalogram (EEG) emotion recognition based on IoT technology is a new research field, which involves emotion induction, EEG feature extraction, and pattern recognition technology. According to the dimensional model of emotion, this paper selects three kinds of Western music fragments, which can express neutral, positive, and negative emotions. It uses these Western music materials to induce an EEG in three emotional states. Comparing the classification effects of the eigenvectors of different rhythms, it is found that the classification with the eigenvectors of the beta and gamma rhythms has the highest accuracy, with the overall average accuracy of 0.842 and 0.841, respectively, and the electrodes that provide features under these two rhythms are in the head. The location distribution of the table was consistent between subjects. Comparing the classification effects of different classifiers, it is found that support vector machine (SVM) and Query-by-Committee (QBC) are better than other classifiers, and the highest average correct rates between subjects on different rhythms are 94.7% and 90.0%, respectively.