Feature selection using support vector machines and bootstrap methods for ventricular fibrillation detection

Alonso‐Atienza, Felipe; Rojo-Álvarez, José Luis; Muñoz, Andrés Pedreño; Vinagre, Juan J.; García‐Alberola, Arcadi; Camps‐Valls, Gustau

doi:10.1016/j.eswa.2011.08.051

Cited by 60 publications

(36 citation statements)

References 45 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Bootstrap resampling is a computer-based method for nonparametric estimation of the distribution of statistical magnitudes, and it can be used to estimate the performance of SVM classifiers [18]. Let V = {(x 1 , y 1 ), .…”

Section: B Bootstrap Resamplingmentioning

confidence: 99%

“…This strategy, however, raises additional requirements to be considered. First, the need of feature selection (FS) techniques to select those relevant and informative parameters in order to increase the efficiency of the learning task, to improve the performance of the detection process, and to better understand how data affect the learning process [18], [21]- [23]. And second, the evaluation and comparison of the proposed algorithms should be assessed over the out of sample test set.…”

Section: Introductionmentioning

confidence: 99%

“…When compared in a standardized way [4], their real performance is reduced from the values presented in the original investigations. The combination of ECG parameters using machine learning techniques, such us neural networks [15]- [17], or support vector machines (SVM) [18]- [20], has been suggested as a useful approach to improve the detection efficiency. This strategy, however, raises additional requirements to be considered.…”

Section: Introductionmentioning

confidence: 99%

See 2 more Smart Citations

Detection of Life-Threatening Arrhythmias Using Feature Selection and Support Vector Machines

Alonso‐Atienza

Morgado

Fernández-Martínez

et al. 2014

IEEE Trans. Biomed. Eng.

Self Cite

187

100

View full text Add to dashboard Cite

Early detection of ventricular fibrillation (VF) and rapid ventricular tachycardia (VT) is crucial for the success of the defibrillation therapy. A wide variety of detection algorithms have been proposed based on temporal, spectral, or complexity parameters extracted from the ECG. However, these algorithms are mostly constructed by considering each parameter individually. In this study, we present a novel life-threatening arrhythmias detection algorithm that combines a number of previously proposed ECG parameters by using support vector machines classifiers. A total of 13 parameters were computed accounting for temporal (morphological), spectral, and complexity features of the ECG signal. A filter-type feature selection (FS) procedure was proposed to analyze the relevance of the computed parameters and how they affect the detection performance. The proposed methodology was evaluated in two different binary detection scenarios: shockable (FV plus VT) versus nonshockable arrhythmias, and VF versus nonVF rhythms, using the information contained in the medical imaging technology database, the Creighton University ventricular tachycardia database, and the ventricular arrhythmia database. sensitivity (SE) and specificity (SP) analysis on the out of sample test data showed values of SE=95%, SP=99%, and SE=92% , SP=97% in the case of shockable and VF scenarios, respectively. Our algorithm was benchmarked against individual detection schemes, significantly improving their performance. Our results demonstrate that the combination of ECG parameters using statistical learning algorithms improves the efficiency for the detection of life-threatening arrhythmias.

show abstract

Section: B Bootstrap Resamplingmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

Detection of Life-Threatening Arrhythmias Using Feature Selection and Support Vector Machines

Alonso‐Atienza

Morgado

Fernández-Martínez

et al. 2014

IEEE Trans. Biomed. Eng.

Self Cite

187

100

View full text Add to dashboard Cite

show abstract

“…The ECG is a diagnostic tool which is widely used in ICUs to monitor and assess patient's heart long-term function(s). A large number of methods aimed for detection of non-invasively measured quantities (such as ECG [2][3][4][5][6][7][8][9][10][11][12][13][14][15][16][17][18][19], Phonocardiogram (PCG) [20] and Arterial Blood Pressure (ABP) [12,17,21] signals) events have yet been proposed based on mathematical models [10], Hilbert Transform (HT) and first derivative [11][12][13], second order derivative [14], wavelet transform and filter banks [5,15,16], soft computing (Neuro-fuzzy, genetic algorithm) [17,18], Hidden Markov Models (HMM) application [19], etc. Based on a comprehensive literature survey among many documented works, it is seen that several features and extraction (selection) methods have been created and implemented by authors in order for detection and classification of ECG waves.…”

Section: Introductionmentioning

confidence: 99%

QRS complex detection based on simple robust 2-D pictorial-geometrical feature

Sabzevari

Moavenian

2013

Journal of Medical Engineering & Technology

View full text Add to dashboard Cite

In this paper a heuristic method aimed for detecting of QRS complexes without any pre-process was developed. All the methods developed in previous studies were used pre-process, the most novelty of this study was suggesting a simple method which did not need any pre-process. Toward this objective, a new simple 2-D geometrical feature space was extracted from the original electrocardiogram (ECG) signal. In this method, a sliding window was moved sample-by-sample on the pre-processed ECG signal. During each forward slide of the analysis window an artificial image was generated from the excerpted segment allocated in the window. Then, a geometrical feature extraction technique based on curve-length and angle of highest point was applied to each image for establishment of an appropriate feature space. Afterwards the K-Nearest Neighbors (KNN), Artificial Neural Network (ANN) and Adaptive Network Fuzzy Inference Systems (ANFIS) were designed and implemented to the ECG signal. The proposed methods were applied to DAY general hospital high resolution holter data. For detection of QRS complex the average values of sensitivity Se = 99.93% and positive predictivity P+ = 99.92% were obtained.

show abstract

“…Consequently, Booster copies similar to a qualifying rule to judge the execution in the FS recipe with the goal that you can assess the difficulty of information attempting to discover arrangement. This paper sees three classifiers: Support Vector Machine, k-Nearest Neighbors recipe, and Naive Bayes classifier [6]. This strategy is rehashed for the k sets of your training test sets, and the prerequisite of the Qmeasurement is registered.…”

Section: Enhanced Modelmentioning

confidence: 99%

Accumulation of High Descriptive Data Classification Using Feature Selection

2017

IJAERD

View full text Add to dashboard Cite

Feature selection using support vector machines and bootstrap methods for ventricular fibrillation detection

Cited by 60 publications

References 45 publications

Detection of Life-Threatening Arrhythmias Using Feature Selection and Support Vector Machines

Detection of Life-Threatening Arrhythmias Using Feature Selection and Support Vector Machines

QRS complex detection based on simple robust 2-D pictorial-geometrical feature

Accumulation of High Descriptive Data Classification Using Feature Selection

Contact Info

Product

Resources

About