What question did this study set out to answer?

This work aims to enhance early detection of cardiovascular diseases using a hybrid machine learning approach.

April 22, 2026Scientific Reports0 citationsOpen Access

IoT-enabled hybrid Slime Mould Algorithm-Support Vector Machine model for early detection of cardiovascular diseases

Q: What is the clinical evidence from this study?

Study design: Other. Population: Cardiovascular disease (n=603). Intervention: IoT-enabled hybrid Slime Mould Algorithm-Support Vector Machine (SMA-SVM) model vs. SVM, DNN, RNN, LSTM, and PSO-CNN models. Primary outcome: Classification accuracy (p=<0.05).

CVC. VenkatesanVokkaligara Sangha Dental College & Hospital TATagrid Abdullah N. AlshalaliPrincess Nourah bint Abdulrahman University VSV. Sivakumar

Key Result

The IoT-enabled hybrid SMA-SVM model achieved 97.77% accuracy in detecting cardiovascular disease, outperforming PSO-CNN and other baseline models.

Key Points

This work aims to enhance early detection of cardiovascular diseases using a hybrid machine learning approach.
Developed an IoT multi-sensor device for data streaming.
Implemented a Slime Mould Algorithm-tuned Support Vector Machine for classification.
Conducted comparative evaluations against multiple established models.
SMA-SVM achieved a 16.11% improvement over SVM in diagnosis.
Outperformed benchmarks including LSTM and DNN by 16.69% and 7.79%, respectively.
Indicated effective real-time clinical decision support for cardiovascular monitoring.

Structured PICO

Does a hybrid Slime Mould Algorithm-Support Vector Machine (SMA-SVM) model improve diagnostic accuracy for cardiovascular diseases compared to standard machine learning algorithms?

Population

603 samples from two public datasets from the UCI repository: SPECTF Heart Dataset (n=267 samples, 44 features) and Arrhythmia Dataset (n=336 samples, 279 features).

Intervention

Slime Mould Algorithm tuned Support Vector Machine (SMA-SVM) with data preprocessing and feature selection via backward elimination.

Comparator

Standard machine learning algorithms including SVM, DNN, RNN, LSTM, PSO-CNN, Logistic Regression (LR), Naive Bayes (NB), and K-Nearest Neighbors (KNN).

Outcome

Diagnostic accuracy, sensitivity, specificity, precision, and F-score for cardiovascular disease detection.

A hybrid SMA-SVM machine learning model demonstrates high accuracy and sensitivity for cardiovascular disease detection, offering a potential tool for IoT-based continuous clinical monitoring.

Main Result

Absolute Event Rate: 97.77% vs 96.08%

p-value: p=<0.05

Limitations

Metaheuristic computational overhead due to successive optimization
Increased training time compared to traditional SVM

Abstract

Early detection is critical to improving outcomes across many diseases, yet cliniciansmust rapidly interpret heterogeneous signals, reports, and images. Automated analysis helps uncover subtle patterns and anomalies that may elude human review. This work targets real-time clinical decision support by streaming data from a portable Internet-of-Things multi sensor device to the cloud and applying a hybrid optimizer classifier pipeline for robust diagnosis. We implemented a Slime Mould Algorithm tuned Support Vector Machine (SMA-SVM) with data preprocessing and feature selection via backward elimination, then partitioned the dataset into training and test sets for evaluation. In comparative experiments, the proposed SMA-SVM outperformed established baselines including SVM, LSTM, DNN, RNN, and PSO-CNN achieving improvements of 16.11%, 16.69%, 7.79%, 11.46%, and 1.75%, respectively, for cardiovascular disease diagnosis. These results indicate that metaheuristic tuning coupled with classical margins can deliver fast, accurate, and resource-efficient predictions suitable for continuous monitoring settings.

Perguntar à IA

Bookmark

View Full Paper