Attention Recognition in Electroencephalogram Signals Based on Multi-feature Fusion

In order to enhance the accuracy of attention detection using EEG(Electroencephalogram) signals, this paper proposes a method based on multi-feature fusion. Due to the complexity of EEG signals, single-feature extraction methods often lack the necessary flexibility and adaptability, making it challenging to comprehensively express the information contained within the signals. To address this issue, the paper presents a multi-feature fusion algorithm based on PE(Permutation Entropy), FE(Fuzzy Entropy), and ME(Mean Energy). Experimental results demonstrate that by utilizing the fused features, the recognition accuracy of two types of signals corresponding to focused attention and relaxation is increased to 90.53%. The ablation experiments and comparisons with existing literature are presented.