Diagnosing the signs of pathological states of a human based on the analysis of heart rate variability

• Published in: 2018 7th Mediterranean Conference on Embedded Computing (MECO) pp. 1 - 4
• Main Authors: Pulavskyi, Anatolii A., Krivenko, Sergey S., Kryvenko, Liudmyla S.
• Format: Conference Proceeding
• Language: English
• Published: IEEE 01.06.2018
• Subjects: artificial intelligence; disease dictionary; Diseases; Electrocardiography; feature engineering; Finite element analysis; heart language; Heart rate variability; internal disease; Medical diagnostic imaging; non-invasive screening; Pathology; SVM; symbolic analysis
• DOI: 10.1109/MECO.2018.8405981

The method of detection of signs of angina pectoris, benign neoplasm of the thyroid gland and peptic ulcer based on the nonlinear analysis of heart rate variability and application of artificial intelligence has been improved. The method works only with time variability of the heart rate and does not use amplitude variability. A comparative analysis of the efficiency of models for signals with a duration of 100, 300, 600 seconds was given. It is shown that the proposed method can work with a single electrocardiogram (ECG) lead and can be integrated into a cloud system. The sensitivity and specificity of diagnosing the signs of angina pectoris for signals with a duration of 300 seconds are 0.81 and 0.92, respectively. The sensitivity and specificity of diagnosing the signs of benign neoplasm of the thyroid gland for signals with a duration of 300 seconds are 0.76 and 0.74, respectively. The sensitivity and specificity of diagnosing the signs of the peptic ulcer for signals with a duration of 300 seconds are 0.80 and 0.71, respectively. Using the posterior probabilities increases the efficiency of the system up to 15%.