Scholarlink

<p>Electrocardiogram (ECG) signal analysis is pivotal in accurately identifying cardiovascular diseases. The proposed work presents an advanced methodology integrating Particle Swarm Optimization (PSO) and Random Forest (RF) algorithms, enhanced with the 8th order IIR Butterworth Filter and Fractional Fourier Transform (FrFT), to improve the analysis and categorization of ECG signals. This approach leverages the 8th order IIR Butterworth Filter to effectively pre-process raw ECG datasets and FrFT's ability to provide a comprehensive time–frequency representation of ECG signals, effectively capturing both frequency and temporal characteristics. PSO is utilized to optimize the parameters of the FrFT, ensuring the most informative features extracted from the ECG data. These features are then used as input into a Random Forest classifier, chosen for its robustness and high performance in handling complex, non-linear data patterns. Extensive experiments are conducted on benchmark ECG datasets to evaluate the performance of the proposed method. The outcomes demonstrate significant improvements in classification accuracy and noise resilience compared to traditional Fourier Transform and wavelet-based methods. The integration of PSO for parameter optimization and RF for classification provides a synergistic effect, enhancing the overall effectiveness of the ECG signal analysis process. The proposed technique was evaluated based on Recall (Re), Precision (Pr), and Accuracy (Acc), yielding results of 99.97% for Re, 99.99% for Pr, and 99.96% for Acc.</p>

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

PSO and RF Based Improved ECG Signal Analysis

  • 期刊论文

  • 发表日期：

    2025年7月6日

查看全文
  • Varun Gupta,
  • Sourav Diwania,
  • Balram Tamrakar,
  • Nikhil Rathi,
  • Jyoti Srivastava

摘要

Electrocardiogram (ECG) signal analysis is pivotal in accurately identifying cardiovascular diseases. The proposed work presents an advanced methodology integrating Particle Swarm Optimization (PSO) and Random Forest (RF) algorithms, enhanced with the 8th order IIR Butterworth Filter and Fractional Fourier Transform (FrFT), to improve the analysis and categorization of ECG signals. This approach leverages the 8th order IIR Butterworth Filter to effectively pre-process raw ECG datasets and FrFT's ability to provide a comprehensive time–frequency representation of ECG signals, effectively capturing both frequency and temporal characteristics. PSO is utilized to optimize the parameters of the FrFT, ensuring the most informative features extracted from the ECG data. These features are then used as input into a Random Forest classifier, chosen for its robustness and high performance in handling complex, non-linear data patterns. Extensive experiments are conducted on benchmark ECG datasets to evaluate the performance of the proposed method. The outcomes demonstrate significant improvements in classification accuracy and noise resilience compared to traditional Fourier Transform and wavelet-based methods. The integration of PSO for parameter optimization and RF for classification provides a synergistic effect, enhancing the overall effectiveness of the ECG signal analysis process. The proposed technique was evaluated based on Recall (Re), Precision (Pr), and Accuracy (Acc), yielding results of 99.97% for Re, 99.99% for Pr, and 99.96% for Acc.