Development of a Diagnostic Tool for Balance Disorders Based on Machine Learning Techniques.

A diagnostic tool is developed for balance disorders based on machine learning techniques. This tool is addressed at experts, in order to support the diagnosis of 5 categories of balance disorders and ultimately 11 specific disorders. Unlike previous works, for each case one general classification model and only one additional specialized classification model are used to provide the recommended diagnosis, while obtaining satisfactory results and overall performance. Certain features are also extracted and identified as determinant for the correct prediction of the diagnostic categories (general classifier) and the diagnoses of each diagnostic category (specialized classifiers).