Development and nursing application of kidney disease prediction models based on machine learning

Area of Science:

• Biomedical Informatics
• Artificial Intelligence in Healthcare
• Nephrology

Background:

• Kidney diseases present challenges in predicting treatment outcomes and disease progression.
• Early detection and accurate prognosis are crucial for effective management of kidney conditions.
• Integrating advanced computational models can enhance diagnostic capabilities in nephrology.

Purpose of the Study:

• To introduce a novel Metaheuristic Red Fox-Optimized Agile Support Vector Machine (MRFO-ASVM) for early kidney disease detection and prognosis.
• To evaluate the effectiveness of the MRFO-ASVM model, incorporating nurses' expertise in data handling.
• To improve the accuracy and reliability of kidney disease diagnosis through advanced machine learning.

Main Methods:

• Data pre-processing using Min-Max normalization.
• Feature extraction performed via Principal Component Analysis (PCA).
• Development and application of the Metaheuristic Red Fox-Optimized Agile Support Vector Machine (MRFO-ASVM) model.

Main Results:

• The MRFO-ASVM model achieved high performance metrics: accuracy (0.92), F1-score (0.67), sensitivity (0.89), precision (0.63), and ROC-AUC (0.99).
• Enhanced parameter performance indicates the model's robustness in kidney disease prediction.
• Nurses' involvement in data collection and analysis contributed to model effectiveness.

Conclusions:

• The MRFO-ASVM model demonstrates significant potential for early kidney disease detection and prognosis.
• Integration into nursing practice can enhance early diagnosis and facilitate personalized patient care.
• This technology advances patient-centered healthcare solutions in nephrology.