Development of a Diagnostic Tool for Balance Disorders Based on Machine Learning Techniques.
Maria Nefeli Nikiforos1, Maria Malakopoulou1, Themis Exarchos2
1Department of Informatics, Ionian University, Corfu, Greece.
Advances in Experimental Medicine and Biology
|January 1, 2022
Summary
A new machine learning tool aids experts in diagnosing balance disorders. It uses specialized models for accurate classification of 11 specific conditions, improving diagnostic performance.
Area of Science:
- Neurology
- Computer Science
- Medical Diagnostics
Background:
- Balance disorders affect patient quality of life and require accurate diagnosis.
- Current diagnostic methods can be complex and time-consuming.
- Machine learning offers potential for improving diagnostic accuracy and efficiency.
Purpose of the Study:
- To develop and validate a machine learning-based diagnostic tool for balance disorders.
- To support expert clinicians in diagnosing 5 general categories and 11 specific balance disorders.
- To identify key features crucial for accurate diagnosis.
Main Methods:
- Development of a general classification model for broad diagnostic categories.
- Implementation of specialized classification models for specific balance disorder diagnoses.
- Extraction and analysis of determinant features for predictive accuracy.
Main Results:
- The diagnostic tool achieved satisfactory results and overall performance.
- The approach utilizing one general and one specialized model proved effective.
- Key features were identified as critical for accurate classification.
Conclusions:
- The developed machine learning tool effectively supports expert diagnosis of balance disorders.
- The dual-model approach enhances diagnostic accuracy for specific conditions.
- Identified determinant features can refine future diagnostic algorithms.
