Journal of Electronic Science and Technology, Volume. 22, Issue 2, 100249(2024)

Efficient ECG classification based on Chi-square distance for arrhythmia detection

Dhiah Al-Shammary1... Mustafa Noaman Kadhim1,*, Ahmed M. Mahdi1, Ayman Ibaida2 and Khandakar Ahmedb2 |Show fewer author(s)
Author Affiliations
  • 1College of Computer Science and Information Technology, University of Al-Qadisiyah, Al Diwaniyah, 58001, Iraq
  • 2Intelligent Technology Innovation Lab, Victoria University, Melbourne, 3011, Australia
  • show less
    Figures & Tables(14)
    Arrhythmia detection based Chi-square classification with the medical organization.
    Schematic representation of the entire methodology, including the preprocessing, feature selection, and the Chi-square-based classifier.
    Results of the preprocessing step (a) before and (b) after normalization.
    Chi-PSO.
    Confusion matrix.
    Confusion matrices of (a) KNN, (b) RF, (c) SVM, (d) NB, (e) DT, and (f) our classifier (Chi-square).
    Confusion matrices of (a) KNN, (b) RF, (c) SVM, (d) NB, (e) DT, and (f) our classifier (Chi-square).
    Evaluated performance of classifiers (a) without and (b) with PSO feature selection.
    • Table 1. Literature overview.

      View table
      View in Article

      Table 1. Literature overview.

      Ref.YearMethodAccuracyLimitation
      [17]2018CNN81.33%The suggested model takes more time to achieve moderate levels of accuracy on the MIT-BIH dataset.
      [18]2016GB + SVM84.82%The performance of the model was evaluated on a redundant dataset including 500 records, exhibiting average accuracy across many classes.
      [19]2019Best first selection (BFS) + RF85.58%A restricted dataset consisting of just 500 entries was used for 16 classes.
      [20]2022RF + SVM77.4%The use of a traditional hybrid ML approach is associated with moderate precision levels and substantial processing costs.
      [21]2021KNN / RF89.83% / 90.21%It takes a lot of data preparation (3-phase preprocessing) to get moderate results.
      [22]2021BiLSTM95%The computational cost is increased due to the prolonged training time of a deep LSTM-BiLSTM model.
      [23]2022MobileNetV2+ BiLSTM91.7%The model needs a longer period of training in order to provide significant outcomes.
      [24]2020DNN94%The use of DNN in conjunction with a genetic algorithm results in a computational process that incurs significant computing expenses.
      [25]2022Fusion of CNN98.8%The model’s ability to perform well across different scenarios is limited due to its dependence on a small dataset.
      [26]2019DWT + Sparse autoencoder (S-AE)96.82%In the presence of elevated levels of noise, the suggested methodology has a limited capability for accurately identifying the precise positions of R-peaks.
      [27]2022SVM + Deep CNN 99.2%The research did not examine feature selection optimization methods, which may include identification of the most relevant deep features, classification performance optimization, and reduction of computational costs.
    • Table 2. Details of classes in the dataset.

      View table
      View in Article

      Table 2. Details of classes in the dataset.

      ClassDescriptionIncluded beatsNumber of extracted records
      NNon-ectopic beatsNormal beats, left bundle branch block, right bundle branch block, nodal (junctional) escape beat, and atrial escape beat100
      SSupraventricular ectopic beatsAberrated atrial premature beat, supraventricular premature beat, atrial premature beat, and nodal (junctional) premature beat100
      VVentricular ectopic beatsVentricular escape beat and premature ventricular contraction100
      FFusion beatsFusion of ventricular and normal beat100
      QUnknown beatsPaced beat, unclassified beat, and fusion of paced and normal beats100
    • Table 3. Details of the implementation environment.

      View table
      View in Article

      Table 3. Details of the implementation environment.

      ComponentsSpecifications
      Processor6th Generation Intel® Core™ i7
      RAM16 GB
      EditorVisual Studio Code
      Programming languagePython 3.12
      Operating systemWindows 10 Pro
    • Table 4. Comparison of our classifier’s performance with that of the standard classifiers without feature selection.

      View table
      View in Article

      Table 4. Comparison of our classifier’s performance with that of the standard classifiers without feature selection.

      ClassifierAccuracy (%)F1-score (%)Precision (%)Recall (%)
      KNN8483.3683.4884
      RF7773.2076.1877
      SVM7872.3483.1478
      NB7474.4975.2274
      DT8384.2987.0183
      Our classifier (Chi-square)8989.4390.4089
    • Table 5. Performance comparison of our classifier (Chi-square) and standard classifiers with feature selection.

      View table
      View in Article

      Table 5. Performance comparison of our classifier (Chi-square) and standard classifiers with feature selection.

      ClassifierAccuracy (%)F1-score (%)Precision (%)Recall (%)
      KNN9696.0696.2396
      RF9392.8992.8393
      SVM9594.9294.8995
      NB8283.2686.9182
      DT9191.1191.2691
      Our classifier (Chi-square)9898.0398.1898
    • Table 6. Comparison of the approach presented in this paper with similar work on the MIT-BIH arrhythmia dataset.

      View table
      View in Article

      Table 6. Comparison of the approach presented in this paper with similar work on the MIT-BIH arrhythmia dataset.

      Ref.YearFeature selectionClassifierAccuracy (%)
      [33]2020Wavelet + Gabor filterBat-rider optimization algorithm deep CNN (BaRoA-DCNN)93.19
      [34]2021Rapid-ramp (RR) + ECG segmentsArtificial deep neural network (ADNN) + Conv1D94.70
      [35]2017Linear discriminant analysis (LDA) + Principal component analysis (PCA) + DWTWeighted k-nearest neighbors (WKNN)96.12
      [19]2019Three-filter feature selection (TFFS)RF + BFS85.58
      [36]2021WaveletCNN97.41
      [37]2020Z-score + High order statistics (HOS) / DWTRF93.45
      KNN72.56
      SVM90.09
      LSTM92.16
      Ensemble SVM94.40
      [38]2019Fractal dimension + Renyi entropy + Fuzzy entropyKNN94.5
      [39]2019HOS + Local binary patterns (LBP) + RREnsemble SVM94.50
      [40]2023ECG segments + RRConv1D MF96.48
      [41]2022RR-intervals + Higher order statistics + DWTEasyEnsemble95.6
      [42]2021Attention mechanismDual-level attentional (DLA) + Convolutional long short-term memory (CLSTM) neural network88.76
      [43]2023Augmented attentionCNN + Attention96.19
      [44]2023Lightweight transformerCNN97.66
      Denoising autoencoder (DAE)97.93
      This workPSOChi-square98
    Tools

    Get Citation

    Copy Citation Text

    Dhiah Al-Shammary, Mustafa Noaman Kadhim, Ahmed M. Mahdi, Ayman Ibaida, Khandakar Ahmedb. Efficient ECG classification based on Chi-square distance for arrhythmia detection[J]. Journal of Electronic Science and Technology, 2024, 22(2): 100249

    Download Citation

    EndNote(RIS)BibTexPlain Text
    Save article for my favorites
    Paper Information

    Category:

    Received: Dec. 20, 2023

    Accepted: Apr. 2, 2024

    Published Online: Aug. 8, 2024

    The Author Email: Kadhim Mustafa Noaman (mustafa.noaman@qu.edu.iq)

    DOI:10.1016/j.jnlest.2024.100249

    Topics