Abstract

This research introduces a novel approach to intelligent IoT-based audio signal processing for healthcare applications. Leveraging advanced feature extraction techniques such as Mel-Frequency Cepstral Coefficients (MFCC) and Wavelet Transform, combined with sophisticated classification models like Convolutional Neural Networks (CNNs) and Support Vector Machines (SVMs), the proposed method demonstrates superior performance in accurately classifying healthcare data. Through extensive experimentation and analysis, the method achieves high accuracy, precision, recall, and F1 score, while exhibiting robustness in discriminating between different classes and maintaining precision in classification, as evidenced by its high AUC-ROC and AUC-PR values. The ablation study provides insights into the significance of key components and parameters, offering guidance for further refinement and optimization of the method. Overall, the proposed method holds promise for revolutionizing healthcare management through proactive monitoring and intervention, leading to improved patient outcomes and healthcare delivery.

Received: August 23, 2023 Revised: November 21, 2023 Accepted: May 29, 2024