3. Cardiac arrhythmia recognition using Stockwell transform and ABC-optimized twin SVM
-
Sandeep Raj
Abstract
The World Health Organization reported that the cardiac disease is a major cause of mortalities worldwide and shall prevail up to the next decade. Therefore, the cardiac healthcare demands an automated computer-aided diagnosis of longer duration electrocardiogram (ECG) signals or heartbeats. In this chapter, an efficient features representation and machine learning methods are combined and developed to process the ECG signals. Initially, the raw heartbeats are preprocessed for eliminating various kinds of noises inherited within them. Consequently, the QRS-wave is located by applying Pan-Tompkins technique within the signals. Following the QRS-wave localization, a rectangular window of fixed size is selected for segmenting the heartbeats. Then, the Stockwell transform method is employed for extracting the time-frequency (TF) information from heartbeats as features. Few coefficients are selected for an efficient representation of heartbeats which further reduces the complexity during processing using classifier. These output coefficients represent the characteristics of individual heartbeats and supports in distinguishing between them based on their morphology. Further, the R-peak to R-peak information between heartbeats are captured and concatenated with the output TF coefficients. As a result, this final feature vector represents each heartbeat that is applied to twin support vector machine classifier to classify these features into their corresponding categories. The classifier performance is also enhanced as its parameters are employed by employing the artificial bee colony algorithm under patient specific scheme. The proposed methodology is validated over PhysioNet database, and the output of the classifier model is compared with the labels of corresponding heartbeats of the database to formulate the results. The experiments conducted reported a higher overall accuracy of 88.22% over existing state-of-the-art methods.
Abstract
The World Health Organization reported that the cardiac disease is a major cause of mortalities worldwide and shall prevail up to the next decade. Therefore, the cardiac healthcare demands an automated computer-aided diagnosis of longer duration electrocardiogram (ECG) signals or heartbeats. In this chapter, an efficient features representation and machine learning methods are combined and developed to process the ECG signals. Initially, the raw heartbeats are preprocessed for eliminating various kinds of noises inherited within them. Consequently, the QRS-wave is located by applying Pan-Tompkins technique within the signals. Following the QRS-wave localization, a rectangular window of fixed size is selected for segmenting the heartbeats. Then, the Stockwell transform method is employed for extracting the time-frequency (TF) information from heartbeats as features. Few coefficients are selected for an efficient representation of heartbeats which further reduces the complexity during processing using classifier. These output coefficients represent the characteristics of individual heartbeats and supports in distinguishing between them based on their morphology. Further, the R-peak to R-peak information between heartbeats are captured and concatenated with the output TF coefficients. As a result, this final feature vector represents each heartbeat that is applied to twin support vector machine classifier to classify these features into their corresponding categories. The classifier performance is also enhanced as its parameters are employed by employing the artificial bee colony algorithm under patient specific scheme. The proposed methodology is validated over PhysioNet database, and the output of the classifier model is compared with the labels of corresponding heartbeats of the database to formulate the results. The experiments conducted reported a higher overall accuracy of 88.22% over existing state-of-the-art methods.
Chapters in this book
- Frontmatter I
- Preface VII
- Contents XI
- List of contributors XIII
- 1. A review of bone tissue engineering for the application of artificial intelligence in cellular adhesion prediction 1
- 2. Implementation and classification of machine learning algorithms in healthcare informatics: approaches, challenges, and future scope 21
- 3. Cardiac arrhythmia recognition using Stockwell transform and ABC-optimized twin SVM 35
- 4. Computational intelligence approach to address the language barrier in healthcare 53
- 5. Recent advancement of machine learning and deep learning in the field of healthcare system 77
- 6. Predicting psychological disorders using machine learning 99
- 7. Automatic analysis of cardiovascular diseases using EMD and support vector machines 131
- 8. Machine learning approach for exploring computational intelligence 153
- 9. Classification of various image fusion algorithms and their performance evaluation metrics 179
- 10. Recommender system in healthcare: an overview 199
- 11. Dense CNN approach for medical diagnosis 217
- 12. Impact of sentiment analysis tools to improve patients’ life in critical diseases 239
- 13. A fuzzy entropy-based multilevel image thresholding using neural network optimization algorithm 253
- 14. Machine learning in healthcare 277
- 15. Computational health informatics using evolutionary-based feature selection 309
- Index 329
Chapters in this book
- Frontmatter I
- Preface VII
- Contents XI
- List of contributors XIII
- 1. A review of bone tissue engineering for the application of artificial intelligence in cellular adhesion prediction 1
- 2. Implementation and classification of machine learning algorithms in healthcare informatics: approaches, challenges, and future scope 21
- 3. Cardiac arrhythmia recognition using Stockwell transform and ABC-optimized twin SVM 35
- 4. Computational intelligence approach to address the language barrier in healthcare 53
- 5. Recent advancement of machine learning and deep learning in the field of healthcare system 77
- 6. Predicting psychological disorders using machine learning 99
- 7. Automatic analysis of cardiovascular diseases using EMD and support vector machines 131
- 8. Machine learning approach for exploring computational intelligence 153
- 9. Classification of various image fusion algorithms and their performance evaluation metrics 179
- 10. Recommender system in healthcare: an overview 199
- 11. Dense CNN approach for medical diagnosis 217
- 12. Impact of sentiment analysis tools to improve patients’ life in critical diseases 239
- 13. A fuzzy entropy-based multilevel image thresholding using neural network optimization algorithm 253
- 14. Machine learning in healthcare 277
- 15. Computational health informatics using evolutionary-based feature selection 309
- Index 329
