Automatic classifier based on heart rate variability to identify fallers among hypertensive subjects. Issue 4 (2nd July 2015)

• Record Type: Journal Article
• Authors: Melillo, Paolo; Jovic, Alan; De Luca, Nicola; Pecchia, Leandro
• Is Part Of: Healthcare technology letters. Volume 2:Issue 4(2015)
• Page Start: 89
• Page End: 94
• Publication Date: 2015-07-02
• Subjects: neurophysiology -- data mining -- electrocardiography -- medical signal processing -- signal classification -- feature selection -- sensitivity analysis -- principal component analysis; automatic classifier -- heart rate variability -- hypertensive subjects -- accidental falls -- high false positive rate -- autonomic nervous system states -- human balance control -- electrocardiogram recordings -- cardiac patients -- linear HRV properties -- nonlinear HRV properties -- data mining -- subject‐based receiver operating characteristic analysis -- RUSBoost -- feature selection method -- principal component analysis -- time 30 min; Biomedical engineering -- Periodicals; Medical technology -- Periodicals; 610.28
• DOI: 10.1049/htl.2015.0012
• Languages: English
• ISSNs: 2053-3713

Abstract

Abstract : Accidental falls are a major problem of later life. Different technologies to predict falls have been investigated, but with limited success, mainly because of low specificity due to a high false positive rate. This Letter presents an automatic classifier based on heart rate variability (HRV) analysis with the goal to identify fallers automatically. HRV was used in this study as it is considered a good estimator of autonomic nervous system (ANS) states, which are responsible, among other things, for human balance control. Nominal 24 h electrocardiogram recordings from 168 cardiac patients (age 72 ± 8 years, 60 female), of which 47 were fallers, were investigated. Linear and nonlinear HRV properties were analysed in 30 min excerpts. Different data mining approaches were adopted and their performances were compared with a subject‐based receiver operating characteristic analysis. The best performance was achieved by a hybrid algorithm, RUSBoost, integrated with feature selection method based on principal component analysis, which achieved satisfactory specificity and accuracy (80 and 72%, respectively), but low sensitivity (51%). These results suggested that ANS states causing falls could be reliably detected, but also that not all the falls were due to ANS states.