Heart disease prediction using IoT based framework and improved deep learning approach: Medical application.

Area of Science:

• Cardiology
• Artificial Intelligence
• Internet of Things (IoT)

Background:

• Heart disease remains a leading global cause of mortality.
• Accurate cardiac disease prediction requires significant medical expertise.
• Internet of Things (IoT) based prediction offers a novel approach using sensor data.

Purpose of the Study:

• To propose an enhanced deep learning framework for accurate heart disease prediction.
• To leverage IoT sensor data for improved disease classification.
• To evaluate the proposed framework's performance against existing techniques.

Main Methods:

• Utilized the Hungarian heart disease dataset collected via IoT sensors.
• Preprocessed data using the Median Studentized Residual approach to handle errors and missing values.
• Feature selection performed using the Harris Hawk Optimization (HHO) algorithm.
• Classified features using a Modified Deep Long Short-Term Memory (MDLSTM) network.
• Optimized LSTM output with the Improved Spotted Hyena Optimization (ISHO) algorithm.

Main Results:

• The implemented framework achieved a high accuracy of 98.01% in heart disease prediction.
• Demonstrated a reduced error rate of 91.11% compared to existing methods.
• Evaluated performance using metrics including specificity, sensitivity, F-Score, Kappa value, Accuracy, BER, and Execution time.

Conclusions:

• The proposed enhanced deep learning framework significantly improves heart disease prediction accuracy.
• The integration of IoT data, HHO, MDLSTM, and ISHO offers a superior approach.
• This methodology represents a promising advancement in non-invasive cardiac disease diagnosis.