A Decisive Metaheuristic Attribute Selector Enabled Combined Unsupervised-Supervised Model for Chronic Disease Risk...

Area of Science:

• Medical Informatics
• Computational Biology
• Data Science

Background:

• Predictive analytics are crucial for assessing chronic disorder risks.
• Traditional attribute selection methods face challenges like complexity and high computational costs.
• Heuristic optimization offers a more efficient approach by reducing computational load and removing irrelevant attributes.

Purpose of the Study:

• To introduce a novel memory-based metaheuristic attribute selection (MMAS) model for optimizing chronic disorder data.
• To enhance the accuracy and efficiency of chronic disease risk prediction.
• To develop a cost-effective decision support system for medical experts.

Main Methods:

• A buffer-enabled heuristic memory-based metaheuristic attribute selection (MMAS) model was developed for local neighborhood search.
• Unsupervised K-means clustering was employed to filter outlier attributes from the data.
• The Naive Bayes classifier was used for the final prediction of chronic disease risks.
• Datasets included heart disease, breast cancer, diabetes, and hepatitis.

Main Results:

• The MMAS model achieved a mean accuracy of 94.5% in predicting chronic disease risks.
• Accuracy decreased to 93.5% when clustering was omitted, highlighting its importance.
• High average precision (96.05%), recall (94.07%), and F-score (95.06%) were recorded.
• The model demonstrated a low latency of 0.8 seconds.

Conclusions:

• Heuristic-based attribute selection combined with clustering and classification significantly improves chronic disease diagnosis.
• The proposed MMAS model offers a more accurate and computationally efficient approach to risk assessment.
• This method can facilitate the development of effective decision support systems for medical professionals.