On the possibilities of the discriminant analysis for the arterial hypertension diagnosis: Evaluation of the short-term heart rate variability feature combinations

Kublanov Vladimir, Dolganov Anton, Hugo Gamboa

Research output: Chapter in Book/Report/Conference proceedingConference contributionpeer-review

1 Citation (Scopus)

Abstract

This paper discusses the possibilities of the Linear and Quadratic Discriminant analysis for diagnosing arterial hypertension patients. For this purpose, electrocardiogram was recorded while performing the tilt testing in two distinct groups: 30 healthy volunteers and 40 patients suffering from the arterial hypertension of II-III degree. Further analysis includes the extraction of 64 features, obtained by statistical, geometric, spectral (Fourier and wavelet) and nonlinear methods. In order to find the best feature combination, a Semi-optimal search of the non-correlated features space is proposed. All calculations were performed in the in-house software written on Python. The results suggest that a 4-features combination of statistical, spectral and nonlinear features provide the most robust classifiers.

Original languageEnglish
Title of host publicationProceedings - 2017 Siberian Symposium on Data Science and Engineering, SSDSE 2017
PublisherInstitute of Electrical and Electronics Engineers Inc.
Pages74-79
Number of pages6
ISBN (Electronic)978-1-5386-1593-5
DOIs
Publication statusPublished - 18 Oct 2017
Event2017 Siberian Symposium on Data Science and Engineering, SSDSE 2017 - Novosibirsk, Akademgorodok, Russian Federation
Duration: 12 Apr 201713 Apr 2017

Conference

Conference2017 Siberian Symposium on Data Science and Engineering, SSDSE 2017
CountryRussian Federation
CityNovosibirsk, Akademgorodok
Period12/04/1713/04/17

Keywords

  • arterial hypertension
  • heart rate variability
  • machine learning
  • python
  • tilt-test

Fingerprint

Dive into the research topics of 'On the possibilities of the discriminant analysis for the arterial hypertension diagnosis: Evaluation of the short-term heart rate variability feature combinations'. Together they form a unique fingerprint.

Cite this