Predicting Heart Failure. Группа авторов

Memory, 9MAGGICMeta-analysis Global Group in Chronic, 24MLMachine Learning, 7MLPMultilayer Perceptron, 8NBNaïve Bayes, 34NLPNatural Language Processing, 16NNNeural Network, 8NRDNationwide Readmissions Database, 21NYHANew York Heart Association, 5RBFRadial Basis Function, 18RDWRed Blood Cell Distribution Width, 26RFRandom Forest, 9RNNRecurrent Neural Network, 8ROTRotation Forest, 16RSRough Set, 8RSARandom Search Algorithm, 10SAESparse Auto-encoder, 9SGDStochastic Gradient Descent, 9SVMSupport Vector Machines, 8

       Chapter 10

AI	Artificial Intelligence, 1 bp Blood Pressure, 11
CAD	Coronary Artery Disease, 15
CT	Computed Tomography, 15
CVDs	Cardiovascular Diseases, 1
DL	Deep Learning, 12
ECG	Electrocardiogram, 1
EIS	Electrochemical Impedance Spectroscopy, 4
IoT	The Internet of Things, 11
MRI	Magnetic Resonance Imaging, 2,13
RI	Refractive Index, 6

       Chapter 11

BMI	Body Mass Index, 9
CHF	Chronic Heart Failures, 2
CIHM	Chronicle Implantable Hemodynamic Monitor, 19
CRT	Cardiac Resynchronization Therapy, 19
HCG	Human Chorionic Gonadotropin, 21
IASD	Inter Atrial Shunt Device, 22
LA	Left Atrial, 17
LAP	Left Atrial Pressure, 18
MCT	Mobile Cardiac Telemetry, 6
NYHA	New York Heart Association, 19
PA	Pulmonary Artery, 17
PAM	Patient Advisory Module, 20
RV	Right Ventricle, 17

       Chapter 12

AI	Artificial Intelligence, 3
ANN	Artificial Neural Networks, 4
AUC	Area Under Curve, 6
CNN	Convolutional Neural Network, 4
CRT	Cardiac Resynchronization Therapy, 6
DL	Deep Learning, 3
DNN	Deep Neural Network, 5
ECG	Electrocardiographic, 5
HF	Heart Failure, 1
k-NN	k-Nearest Neighbors, 5
LVAD	Left Ventricular Assist Device, 7
ML	Machine Learning, 3
PPGs	Photoplethysmograms, 8
RF	Random Forest, 4
RNN	Recurrent Neural Network, 5
RV	Right Ventricular, 2
RVF	Right-ventricular Failure, 7
RVFRS	Right Ventricular Failure Risk Score, 7
SVM	Support Vector Machine, 4

       Chapter 13

ABP	Arterial Blood Pressure, 4
AF	Atrial Fibrillation, 11
CAD	Coronary Artery Diseases, 19
CardioMEMS	Cardio-Microelectromechanical system, 3
CRT	Cardiac Resynchronization Therapy, 18
CRT-D	Cardiac Resynchronization Therapy Defibrillator, 4
CVDs	Cardiovascular Diseases, 3
ECG	Electrocardiogram, 2
FDA	Food and Drug Administration, 13
HF	Heart Failure, 1
hour	Heart Rate, 13
ICDs	Implantable Cardioverter-defibrillators, 5
LAP	Left Atrial Pressure, 5
LVAD	Left Ventricular Assist Device, 18
MI	Myocardial Infarction, 14
NSTEMI	Non-ST-elevation Myocardial Infarction, 18
NYHA	New York Heart Association, 10
OHRM	Optical Heart Rate Monitor, 9
PAP	Pulmonary Artery Pressure, 3
PD	Photodiode, 9
PPG	Photoplethysmogram, 4
RR	Respiration Rate, 13
STEMI	ST-elevation Myocardial Infarction, 18
SVM	Support Vector Machine, 18
VF	Ventricular Fibrillation, 11

Acknowledgment

This publication was supported by Qatar University Internal Grant No. IRCC-2020-013 and Sultan Qaboos University through Grant # CL/SQU-QU/ENG/20/01, respectively. The findings achieved herein are solely the responsibility of the authors.

1 Invasive, Non-Invasive, Machine Learning, and Artificial Intelligence Based Methods for Prediction of Heart Failure

       Hidayet Takcı

      1.1 Introduction

      Heart diseases are the deadliest in the world. Of the many diseases included in the category of heart disease, the most prominent is coronary artery disease (CAD), which causes heart attacks. CAD, high blood pressure, and many other heart diseases cause heart failure (HF). With HF being a consequence of heart disease, the prediction of HF is related to the prediction of diseases categorized as heart disease.

      In this chapter, the diagnosis of HF is discussed in terms of invasive/non-invasive and artificial intelligence/machine learning techniques. Invasive and non-invasive techniques are distinct in the way the patient is treated. Invasive methods are usually associated with a physical intervention in the body. This intervention involves operations such as taking blood for blood analysis and not pressing strongly on the abdominal area. Non-invasive methods include physical therapy, taking blood pressure, and temperature measurement. In today’s world, where information technologies have evolved in every field, the field of health has also received its share. Computer-aided clinical decision support systems provide the strongest support for diagnostic studies today. The most important components of computer-aided diagnosis are artificial intelligence and machine learning systems that offer a wide range of services from smart assistant applications to imaging techniques. Artificial intelligence and machine learning have a healing role in electrocardiography, echocardiography, and similar invasive and non-invasive techniques.

In the second section of this chapter, HF will be described and its causes, symptoms, and treatment revealed. The third section will explain diagnosis by invasive and non-invasive methods. Computer-aided diagnosis and decision support systems are briefly mentioned in the fourth section. The fifth section looks first at what artificial intelligence is and then presents fields and examples of artificial intelligence supported studies. The sixth section explains machine learning, learning types, machine learning algorithms, and machine learning based diagnostic studies. The chapter’s concluding section summarizes studies and comments upon diagnostic studies for HF.

      1.2 Heart Failure

      HF belongs to a class of diseases that occur due to several diseases known as heart diseases and can be fatal if left untreated. In this section, HF is defined and the factors causing it, its symptoms, and its treatment are examined.

      1.2.1 What is HF?

      HF is sometimes known as congestive HF (CHF). HF is not a failure of the heart, but a condition in which the heart muscle cannot pump enough blood. It occurs due to, for example, narrowing of the coronary arteries, cardiovascular diseases, and high blood pressure.

      HF can occur at any age, but is more common in older people. Although it is not possible to cure the disease, some of its symptoms can be improved by the correct interventions. Conversely, faulty interventions will make the situation worse than before.

      The most common symptoms of HF are shortness of breath after movement or during rest, fatigue and weakness, swelling