Sleep staging method based on improved particle swarm algorithm and twin support vector machine

Abstract

The invention relates to a sleep staging method based on an improved particle swarm algorithm and a twin support vector machine. The method comprises: on the basis of a particle swarm search algorithmand a twin support vector machine, introducing an adaptive gain feedback rate and gradient random disturbance to the particle swarm search algorithm, and combining an improved particle swarm optimization algorithm with the twin support vector machine to form an efficient and high-precision classifier for adaptive parameter adjustment; classifying and recognizing different sleep states in the whole night sleep process by taking the energy change of each frequency band of the electroencephalogram signal as a characteristic. According to the classifier technology, a self-adaptive parameter setting strategy is adopted, hyper-parameters do not need to be set manually, experience errors possibly caused by manual hyper-parameter determination are avoided, local optimal parameters cannot occur, and the data classification accuracy is guaranteed.

Description

technical field [0001] The invention relates to the field of computer technology, in particular to a sleep staging method based on an improved particle swarm algorithm and a twin support vector machine. Background technique [0002] Sleep is a necessary physiological activity for the human body, manifested as a reduction in the body's ability to respond to external stimuli and a temporary interruption of consciousness. In the waking state, humans and animals can carry out a series of activities; while in the sleeping state, the human body and brain will be fully relaxed, enabling memory consolidation, body recovery and integration. Under normal conditions, based on the average sleep time of 8 hours per day, about one-third of the human life is spent in sleep activities. Therefore, there is a very close relationship between sleep and people's life and work, and its importance is second only to breathing and heartbeat. The quality of sleep directly or indirectly affects peopl...

AI Technical Summary

Problems solved by technology

Under normal conditions, people sleep for 7-8 hours throughout the night, and the amount of sleep characteristic data generated is very large. It is time-consuming and laborious to manually mark the sleep data; and the results of manual interpretation are subject to great subjectivity. Easily affected by factors such as physician experience and environment, leading to inaccurate classification results

Images