International Journal of Advances in Engineering & Scientific Research
Print ISSN : 2349 –4824
Online ISSN : 2349 –3607
Frequency : Continuous
Current Issue : Volume 12 , Issue 1
2025
Mr. K. Anguraju, Muhammad Yahya S, Sri Vinayakamoorthy M, Vishnu R
Mr. K. Anguraju, Muhammad Yahya S, Sri Vinayakamoorthy M, Vishnu R, Department of Computer Science and Engineering, Kongunadu College of Engineering and Technology, Tamilnadu, India
DOI : https://doi.org/10.5281/zenodo.14928824 Page No : 138-150
Published Online : 2025-02-26
Download Full Article : PDF Check for Updates
Peripheral Artery Disease (PAD) often coexists with other cardiovascular conditions, necessitating accurate comorbidity assessment. This study presents an intelligent self-organizing Long Short-Term Memory (LSTM) model to classify and assess PAD comorbidities using radial artery pulse wave data. A case-control study was conducted with 202 healthy individuals, 187 patients with CoronaryArteryDisease (CAD), and 73 patients with Heart Failure (HF), sourced from Shanghai Municipal Hospitals of Traditional Chinese Medicine. Pulse wave data were recorded, denoised, and standardized using min-max normalization.Four deep learning models—Bidirectional LSTM (Bi-LSTM), Convolutional Neural Network (CNN), Gated Recurrent Units (GRU), and LSTM—were employed to classify pulse wave characteristics. Data imbalance was addressed using Synthetic Minority Oversampling Technique (SMOTE), improving minority class representation. Model performance was evaluated using accuracy, precision, recall, F1-score, specificity, and AUC metrics.The results demonstrated that SMOTE-enhanced data distribution improved classification performance. Additionally, t-distributed Stochastic Neighbor Embedding (t-SNE) was applied to visualize high-dimensional data relationships. The proposed LSTM-based model exhibited superior long-term dependency learning, offering a robust approach for PAD comorbidity assessment.
Keywords: Deep Learning, Pulse Wave Analysis, Synthetic Minority Oversampling Technique (SMOTE), t-Distributed Stochastic Neighbor Embedding (t-SNE) and Comorbidity Assessment
