An automatic sleep staging method based on dual feature screening

Abstract

An automatic sleep staging method based on dual feature selection, first extracting the two-lead sleep EEG signal and one-lead horizontal oculoelectric signal of the subject, and filtering the original EEG and oculoelectric signal; Multiple features are extracted from EEG and EoG signals. Then, the dual feature screening method combined with the Fisher scoring method and the sequential forward method is used to select the optimal feature subset; the dual feature screening not only greatly reduces the feature dimension, but also reduces the redundancy between features; finally, through Use the support vector machine classifier to identify the optimized features and complete the automatic sleep staging; this method can well improve the objectivity, accuracy and simplicity of automatic sleep staging, with high accuracy, low computational complexity, and easy operation , Easy to promote, and can obtain considerable social and economic benefits.

Description

technical field [0001] The invention relates to the field of automatic sleep staging, in particular to an automatic sleep staging method based on double feature screening. Background technique [0002] Sleep is a physiological process generally required by all higher organisms, including humans. People spend one-third of their time sleeping, and long-term lack of sleep will cause serious consequences. Long-term lack of sleep will not only endanger people's health, but also have many negative effects on life and society. Therefore, sleep research is receiving more and more attention. Sleep staging is the basis of sleep research and plays a key role in sleep research. It can objectively evaluate sleep quality, provide clinical basis for the treatment and diagnosis of sleep-related diseases, and provide clinical basis for the diagnosis and research of neurological diseases such as epilepsy and dementia. [0003] According to the latest standard of sleep classification of the...

AI Technical Summary

Problems solved by technology

The Fisher score method is to calculate the score of each feature separately according to the Fisher criterion, which ignores the impact of the feature combination on the classification; while the sequential forward method is a heuristic search strategy, and the optimal feature subset obtained by it may only be some The feature combination with a very common individual contribution rate ignores the feature with the largest individual contribution rate, and the calculation process takes a long time

Images