Quantum-inspired seagull optimised deep belief network approach for cardiovascular disease prediction

Area of Science:

• Cardiovascular Disease Research
• Artificial Intelligence in Healthcare
• Computational Biology

Background:

• Early detection of cardiovascular diseases is critical for reducing mortality.
• Existing diagnostic models often face challenges with accuracy and efficiency.
• Advanced machine learning techniques are needed to enhance cardiovascular disorder identification.

Purpose of the Study:

• To propose a novel hybrid model for improved cardiovascular disease prediction.
• To enhance the accuracy and efficiency of cardiovascular disorder identification using artificial intelligence.
• To leverage quantum-inspired optimization for better machine learning model performance.

Main Methods:

• Data preprocessing: cleaning, transformation, and standardization of the Cleveland Heart Disease dataset.
• Model development: combining a deep belief network (DBN) with a quantum-inspired seagull optimization algorithm (QISOA) for weight and bias optimization.
• Performance evaluation: comparing the QISOA-DBN model against traditional machine learning and metaheuristic algorithms.

Main Results:

• The QISOA-optimized DBN model achieved a high accuracy of 98.6%.
• The model demonstrated superior precision (97.6%), recall (96.8%), and F1-score (97.1%).
• Outperformed traditional machine learning, metaheuristic algorithms, and state-of-the-art methods in cardiovascular disease prediction.

Conclusions:

• The proposed QISOA-DBN model significantly improves cardiovascular disease prediction accuracy.
• Quantum-inspired optimization enhances DBN's feature learning and classification capabilities.
• This hybrid approach offers a promising tool for early and accurate diagnosis of cardiovascular disorders.