Biomarkers.

Area of Science:

• Neurology
• Artificial Intelligence
• Biomedical Informatics

Background:

• Dementia affects over 50 million globally, with numbers projected to triple by 2050.
• Early and accurate dementia prediction is crucial for effective patient management.
• This study explores advanced machine learning for enhanced predictive accuracy.

Purpose of the Study:

• To investigate the predictive capabilities of Support Vector Regression (SVR), Gradient Boosting Machine (GBM), and Adaptive Neuro-Fuzzy Inference System (ANFIS) for dementia diagnosis.
• To compare the performance of these machine learning models using key evaluation metrics.
• To assess the potential of hybrid and ensemble methods in improving dementia prediction.

Main Methods:

• Utilized a dataset of 149 participants (aged 60-96) with nine clinical and imaging biomarkers.
• Employed Support Vector Regression (SVR), Gradient Boosting Machine (GBM), and Adaptive Neuro-Fuzzy Inference System (ANFIS) models.
• Applied 70% for training and 30% for validation, evaluating with R², RMSE, and MSE metrics.

Main Results:

• All investigated machine learning models demonstrated accurate dementia prediction.
• Adaptive Neuro-Fuzzy Inference System (ANFIS) outperformed SVR and GBM in precision and consistency.
• ANFIS achieved a perfect R² of 1.0 in both training and testing, while GBM and SVR showed high accuracy (R² of 0.998 and 0.995).

Conclusions:

• Ensemble and hybrid machine learning approaches significantly enhance dementia prediction accuracy.
• Findings support the efficacy of ANFIS for reliable and early dementia diagnosis.
• This research paves the way for improved diagnostic tools in dementia care.