Research On The Correlation Between Age And Heart Rate Variability

Heart Rate Variability(HRV)refers to the phenomenon that sinus heart rate changes periodically within a certain period of time.As age grows,HRV will change.To clarify the internal relationship between HRV and age is of great reference from human health evaluation and clinical diagnosis of cardiovascular diseases.This article takes HRV characteristics as the research object,and focuses on the correlation and degree of correlation between HRV characteristics and age,and then applies the correlation analysis results to age evaluation.In the process of HRV feature extraction,due to the influence of environment or individual factors,ectopic beating points that do not meet the normal heart rate beat will appear,called dead points,leading to deviations in the calculation of HRV features.In response to this problem,this article creatively proposes a dead pixel detection and replacement method based on box plots in the preprocessing stage of the heart rate variability curve.This method realizes the automatic detection and replacement of dead pixels.Compared with the traditional dead pixel removal method,this method ensures the dependence on the sequence of time.In order to clearly to analyze the correlation between age and HRV characteristics,this article is based on statistical analysis,using Spearman's correlation coefficient analysis and Kruskal-Wallis significance test method to conduct a comprehensive analysis from two perspectives.The relationship between 29 HRV features and age was analyzed on four age groups.According to the experimental results,it was determined that 23 of the extracted features were correlated with age,and the HRV features and age was obtained.Sort by the degree of relevance.In order to improve the speed of feature selection,this paper applies the results of correlation analysis to feature selection based on the idea of sequence forward selection algorithm,proposes a feature selection strategy based on correlation degree ranking,and verifies the effectiveness of the selected feature sets in SVM,KNN and decision tree.Experimental results show that the algorithm can effectively reduce the running time of the algorithm and improve the efficiency.The accuracy increased by from 13.9percent,9.5 percent and 8.4 percent,respectively.