Weight optimized neural network for heart disease prediction using hybrid lion plus particle swarm algorithm

Area of Science:

• Cardiology and Clinical Data Analysis
• Machine Learning in Healthcare
• Computational Intelligence

Background:

• Heart disease is a leading cause of mortality and morbidity globally.
• Accurate prediction of heart disease is crucial for effective clinical management.
• Increasing healthcare data complexity challenges traditional analysis and prediction models.

Purpose of the Study:

• To propose a novel heart disease prediction model using machine learning.
• To enhance prediction accuracy through advanced feature extraction and dimensionality reduction.
• To optimize the prediction model using a hybrid meta-heuristic algorithm.

Main Methods:

• Feature extraction using statistical and higher-order statistical methods.
• Dimensionality reduction via Principal Component Analysis (PCA).
• Heart disease prediction using a Neural Network (NN) optimized by a hybrid Particle Swarm Optimization (PSO) and Lion Algorithm (LA) update (PM-LU) approach.

Main Results:

• The proposed PM-LU-NN model demonstrated superior accuracy compared to conventional methods.
• Accuracy improvements were 3.85% over LM-NN and PSO-NN, 12.5% over WOA-NN and FF-NN, and 7.41% over LA-NN.
• The hybrid algorithm effectively addressed the optimization challenges in NN weight tuning.

Conclusions:

• The PM-LU-NN model offers a significant advancement in heart disease prediction accuracy.
• Hybrid meta-heuristic optimization is effective for enhancing machine learning-based clinical prediction models.
• This approach can aid in managing complex e-healthcare data for improved patient care.