TABLE 1.

Summary of previous works and their advantages and disadvantages.

Year	Author	Title	Algorithm	Dataset	Advantages	Disadvantages
2008	Babak Mohammadzadeh Asl, Seyed Kamaledin Setarehdan and Maryam Mohebbi	Support vector machine‐based arrhythmia classification using reduced features of heart rate variability signal	Algorithm based on Generalized Discriminant Analysis (GDA) Feature Reduction and Support Vector Machine (SVM) Classifier	MIT‐BIH	1. Diagnosis with high accuracy. 2. Reduction of processing time and possibly creating an online arrhythmia detection system. 3. Reducing the number of features increases the accuracy of decision‐making	1. Inability to recognize arrhythmias such as left bundle branch block and right bundle branch block. 2. Dependence on specific HRV signals. 3. Loss of important information in the ECG signal
2010	Zine‐Eddine Hadj Slimane and Amine Naït‐Ali	QRS complex detection using Empirical Mode Decomposition	Empirical Mode Decomposition Algorithm	MIT‐BIH	1. Multiple steps with high accuracy. 2. Significant improvement in QRS complex detection	1. Computational time complexity leads to slowness in processing. 2. Dependence on specific MIT‐BIH datasets. 3. Complexity of algorithm implementation
2012	Yakup Kutlu and Damla Kuntalp	Feature extraction for ECG heartbeats using higher order statistics of WPD coefficients	1. Higher order statistics (HOS) of wavelet packet decomposition (WPD) coefficients 2. KNN	MIT‐BIH	1. High accuracy in heart rate detection. 2. WPD noise removal has a higher frequency resolution. 3. Strong performance in classification with KNN	1. Computational complexity in using higher order statistics calculation. 2. Poor performance in diagnosing other types of cardiac disorders. 3. Low detection accuracy in unusual conditions. 4. Reduction of KNN performance in large and complex data
2016	Sandeep Raj, Kailash Chandra Ray, and Om Shankar	Cardiac arrhythmia beat classification using DOST and PSO‐tuned SVM	1. Discrete orthogonal Stockwell transform (DOST) 2. Particle swarm optimization (PSO) 3. The support vector machine (SVM) classifier	MIT‐BIH	1. Improving the accuracy of diagnosis and reducing mistakes. 2. Better diagnosis in the face of dynamic changes	1. Computational complexity and the need for high processing power. 2. Increasing execution time in parameter optimization with PSO
2017	Santanu Sahoo, Bhupen Kanungo, Suresh Behera and Sukanta Sabut	Multiresolution wavelet transform‐based feature extraction and ECG classification to detect cardiac abnormalities	1. Multiresolution wavelet transform 2. Neural network (NN) 3. Support vector machines (SVM) classifiers	MIT‐BIH	1. High accuracy in diagnosis and classification. 2. Noise removal	Increase computational complexity using wavelet transform
2019	Fakheraldin Y. O. Abdalla, Longwen Wu, Hikmat Ullah, Guanghui Ren, Alam Noor and Yaqin Zhao	ECG arrhythmia classification using artificial intelligence and nonlinear and nonstationary decomposition	Nonlinearity and nonstationary decomposition methods using complete ensemble empirical mode decomposition with adaptive noise (CEEMDAN)	MIT‐BIH	1. Extracting special features to diagnose cardiac irregularities. 2. High accuracy in diagnosis and classification	It has computational complexity due to the use of nonlinear analysis methods and the extraction of particular features
2020	Hadjer Zairi, Malika Kedir Talha, Karim Meddah and Saliha Ould Slimane	FPGA‐based system for artificial neural network arrhythmia classification	1. Field programmable gate array (FPGA) 2. Wavelet transform 3. Multilayer perception (MLP)	MIT‐BIH	1. High accuracy and generalizability. 2. Minimizing features and reducing energy consumption. 3. Fast and real‐time implementation 4. Long‐term and interactive monitoring. 5. Low error rate in data classification	1. Complexity of FPGA implementation. 2. Resource limitations. 3. Using an FPGA chip is expensive and time‐consuming. 4. Limitation in the diagnosis of some diseases. 5. In some cases, there is a decrease in classification accuracy and errors in pattern recognition
2020	Varun Gupta and Monika Mittal	Arrhythmia detection in ECG signal Using fractional wavelet transform with principal component analysis	1. Fractional wavelet transform (FrWT) 2. Yule‐Walker autoregressive modeling (YWARM) 3. Principal component analysis (PCA)	MIT‐BIH	1. Noise cleaning. 2. Improve detection accuracy	1. More time is needed for processing due to the higher computational complexity of combining methods. 2. Due to the use of variance estimation with PCA, there is a possibility of less accuracy in diagnosing some cardiac irregularities. 3. Optimizing the method's performance requires better configuration and parameter setting