SAHS automatic identification method based on multiple baseline features of SaO2 signals

Abstract

The invention discloses an SAHS automatic identification method based on multiple baseline features of SaO2 signals, and the method comprises the steps: S1, collecting the SaO2 signals in an SHHS database, and carrying out the preprocessing of the SaO2 signals; S2, extracting a plurality of basic data features, and screening out an optimal feature; S3, extracting a plurality of custom baseline related features and combining the custom baseline related features with the optimal feature to form a feature data set; S4, calculating an AHI value according to the annotation file corresponding to the data extracted from the SHHS database, classifying the feature data set according to the AHI value, and selecting a random equilibrium data method to process the unbalanced data set to obtain an equilibrium data set; and S5, taking the balanced data set as the input of a random forest classifier, and training and testing the data set to obtain a final classification result. According to the invention, the SaO2 signal change condition is better reflected, and the illness severity of a subject is better reflected through baseline related characteristics; and a complete balanced data set is combined through a random balanced data method, so that the data randomness is ensured, and the final classification result is more accurate.

Description

technical field [0001] The invention belongs to the technical field of biomedicine, in particular to a SaO-based 2 A method for the automatic identification of SAHS of multiple baseline characteristics of the signal. Background technique [0002] Sleep apnea-hypopnea syndrome (SAHS) is the most common sleep-related respiratory disease. SAHS causes partial or complete cessation of breathing during sleep, known as hypopnea and apnea events, respectively. However, because SAHS patients mainly manifest symptoms such as snoring and suffocation during sleep, many people do not realize that they have SAHS. [0003] Subjects used portable devices to monitor their nighttime sleep SaO 2 By monitoring and analyzing the signals, not only can you know your own illness in time, but you can also save the hospital's cumbersome inspection steps and inspection costs. [0004] However, currently based on SaO 2 There are still some problems in signal detection and analysis, such as normal ...

AI Technical Summary

Problems solved by technology

[0004] However, currently based on SaO 2 There are still some problems in signal detection and analysis, such as normal sleep state, night sleep SaO 2 The signal will also fluctuate within the normal range, so only extracting the time domain features, frequency domain features and nonlinear features of the signal cannot accurately identify SAHS

Another example is that there is an obvious data imbalance in the Sleep Heart Health Study (SHHS), which also leads to the inability to accurately identify SAHS in the end.

Images