Review Paper on Cardiac Arrhythmia using Different Machine Learning Algorithms

Abstract

cardiac arrhythmia is a heart condition in which there will be irregularity of the heart beat. This project aims to detect and classify arrhythmia into 14 different variants. Mainly there are two types of cardiac Arrythmia-Tachycardia in which heart rhythm will beat with a rate of more than 100 beats per minute. Bradycardia in which heart rhythm will be slow with a rate below 60 beats per minute. IN this work we use the several machine algorithms like SVM , random forest many other algorithms ,we will try get the highest accuracy by using different machine learning tools in three steps we can do the prediction of the cardiac arrhythmia, we will train the data by using different machine learning algorithms In the second stage upload the model into our Arduino and start predicting real ECG signal which get from the ECG sensor we are using, then the proposed system will send the signal and the prediction results to the mobile application and it stores it on the cloud for the easy access by the user or doctor

Introduction

I. INTRODUCTION

One of the common disease of heart is cardiac Arrhythmia. There are different types of heart arrhythmia such as atrial  fibrillation,  ventricular  escape  and  ventricular fibrillation may cause heart strokes and cardiac arrest which results in death of the person .  An arrhythmia occurs  when  there  are  some  disorders  in  the  heart . Different  arrhythmias  provide  different  ECG  patterns which differs from person to person .

The arrhythmias such  as ventricular  as  well as atrial fibrillations and flutters are life-threatening and may lead to the heart stroke and may also lead to sudden death of the person. Possibilities of cardiac Arrhythmia is high for a patient who had previously suffered from a heart attack.

Heart disease results in death of nearly 38,000 people annually. These problems can be solved by using the appropriate different machine learning techniques. In this project we use the ECG signal and different parameters both in real time and offline dataset and we train the model and then the model predicts whether the human is has high chance or low chance of getting the  cardiac Arrhythmia

II. LITERATURE SURVEY

A. Classification of cardiac arrhythmia by halil Ibrahim BULBUL

In study conducted by ECG signals Support Vector Machine  (SVM)  and  Multilayer  Sensor  (MLP) classifiers were used because the MLP and SVM classifiers gave the most successful results when working in this area. The performance of the classifiers to be used is compared according to the time and other performance criteria. The main aim of this study is to apply some wave transformation techniques such as DWT, CWT, DCT to ECG signals in order to improve the classification and also to increase the performance of these wave transformations.

B. Arrhythmia Detection using Amplitude Difference Features  Based  on  Random  Forest published by Juyoung Park

This  paper  point  was  to  add  to  the  finding  of arrhythmia by presenting another element called adequacy contrast to heartbeat grouping in light of two cycles: 1) heartbeat identification and element extraction; and 2) irregular timberland classifier to group pulses by their elements.

Board analyses researching the impacts of adding another component in heartbeat grouping utilizing the MIT-BIH arrhythmia information base show that considering an abundancy distinction component  can work on the exhibition of heartbeat characterization by diminishing misleading positive and bogus negative rates.

C. Cardiovascular disease risk score prediction models for women published by Louise GH Goh

This paper proposes the Information on socioeconomics, ailments and treatment, smoking conduct, dietary conduct, and exercise designs were gathered. Physical estimations were likewise taken. The 10-year risk was determined utilizing the Framingham model, SCORE (Systematic Coronary Risk Evaluation) risk diagram for generally safe and high gamble locales, the overall CVD, and worked on broad CVD risk score models in 4,354 females matured 20-69 years with no coronary illness, diabetes, or stroke at standard from the third.

D. Chest pain in the emergency room: value of the HEART score published by ALBERT JACOB SIX

This paper exhibits that Chest torment is one of the most well-known reasons for show to the trauma center. The determination of non-ST-rise intense coronary disorder regularly causes vulnerability. Old style contemplations for risk separation are History, ECG, Age, Risk factors and Troponin (HEART). Each can be scored with nothing, a couple of focuses, contingent upon the degree of the anomaly. The HEART score is the amount of these five contemplations.

III. ALGORITHM

IV. PROPOSED METHODOLOGY

1. The ECG paddles are placed on the chest or wrist of the person. The ECG sensors and heart beat sensors are connected to Arduino as input and the Transmitting and Receiving part of Arduino is connected to transmitter and receiver of
2. Wi-Fi module from there the collected data is then uploaded to things speak server.
3. The distance between the heart beat is measured and if there is any abnormal difference between the distance of heartbeats then message will be sent to the user via telegram so that the user can take necessary precautions by consulting a doctor. This is regarding the hardware part
4. When it comes the software part the different parameters like blood pressure , age, gender and sugar levels of the person is taken in to consideration and then data sets of these parameters are downloaded from Kaggle then it is compared with user data by applying machine learning algorithms cardiac arrhythmia is predicted

V. OBJECTIVES

A. Surveying the patients who are suffering from Arrhythmia and collecting the certain data samples to overcome the further problems faced by the patients in future level.

B. Collecting different sets of data related to the variants in Arrhythmia and proposing the model to alert the person or care-taker at critical conditions.

C. Implementing a Light weight Arrhythmia model, Which helps in monitoring the person condition time to time and It sends the certain ECG signal to the care-taker of that particular person.

D. Using specific software and hardware components related to the proposed model ,We able to extract accurate and précised ECG signal of particular person.

Conclusion

A. The accuracy in the medical field is very important. B. There are many different machine learning algorithms C. In our project we are implementing by considering combination of SVM and random forest. D. By using these algorithms it is expected to get an accuracy of more than 98%. E. By this we can predict and detect Weather the person gets Arrhytmia or not

References

[1] CLASSIFICATION OF ECG ARRHYTHMIA WITH MACHINE LEARNING TECHNIQUES https://ieeexplore.ieee.org/document/8260688 [2] Arrhythmia Detection using Amplitude Difference Features Based on Random Forest https://ieeexplore.ieee.org/document/7319561 [3] Cardiovascular disease risk score prediction models for women and its applicability to Asians https://www.researchgate.net/publication/26095152 9_Cardiovascular_disease_risk_score_prediction_m odels_for_women_and_its_applicability_to_Asians [4] Chest pain in the emergency room: value of the HEART score https://www.researchgate.net/publication/23135754 _Chest_pain_in_the_emergency_room_Value_of_th e_HEART_score

Copyright

Copyright © 2022 Harshith K Y, Moksith Reddy, Chandana H S, Tanuja S, Mr. Vivek Raj. This is an open access article distributed under the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.