EEG Denoising Attention Classification Method Based on SVMD Coupled with Otsu's Method Improved Fluctuation Discrete Entropy Threshold

Attention assessment has a key role in the diagnosis of attention deficit hyperactivity disorder (ADHD). In this paper, we propose a group modelling classification method based on single-channel EEG for distinguishing attentive and inattentive states. First, the original signal is decomposed into multiple Intrinsic Mode Functions (IMFs) using Successive Variational Mode Decomposition (SVMD). Then, pure IMFs and noisy IMFs are adaptively filtered by combining the fluctuation-based dispersion entropy improved by Otsu method (OFDE), and the pure IMFs are reconstructed to obtain the denoised signal. After extracting the statistical properties and time frequency features of the denoised signals, classification is performed using Random Forest (RF) and Support Vector Machine (SVM). Numerical simulation results show that the classification method based on SVMD-OFDE proposed in this paper has a significant improvement in classification accuracy compared to using only SVMD or the traditional method combined with OFDE.