Automatic selection model to identify neurodegenerative diseases.

Area of Science:

• Neurology
• Biomedical Engineering
• Data Science

Background:

• Parkinson's disease and Huntington's disease are debilitating neurological disorders.
• Accurate and early diagnosis is crucial for effective management.
• Non-invasive diagnostic methods are highly desirable to reduce patient burden.

Purpose of the Study:

• To evaluate the effectiveness of machine learning algorithms in classifying Parkinson's disease and Huntington's disease.
• To utilize non-invasive biomarker data for disease classification.
• To explore the potential of a unique Mexican human gait biomarker database.

Main Methods:

• Collected accelerometer biomarker data (x, y, z values) from Parkinson's disease patients, Huntington's disease patients, and healthy controls.
• Implemented an automatic selection model for disease classification.
• Employed Random Forest, Random Subspace Method, and K-star algorithms with automated parameter optimization.

Main Results:

• Achieved a 0.893 precision rate for classifying Parkinson's disease and Huntington's disease using the Random Subspace Method.
• Demonstrated the efficacy of the automatic selection model method.
• Highlighted the potential of machine learning in medical diagnosis.

Conclusions:

• Machine learning algorithms, particularly the Random Subspace Method, show significant potential for classifying neurological disorders.
• Non-invasive biomarker data can be effectively utilized for disease diagnosis.
• This research contributes to advancing non-invasive diagnostic approaches in neurology and underscores the role of AI in healthcare.