Classification of Stabilometric Time-Series Using an Adaptive Fuzzy Inference Neural Network System

Abstract

Stabilometry is a branch of medicine that studies balance-related human functions. The analysis of stabilometric-generated time series can be very useful to the diagnosis and treatment balance-related dysfunctions such as dizziness. In stabilometry, the key nuggets of information in a time series signal are concentrated within definite time periods known as events. In this study, a feature extraction scheme has been developed to identify and characterise the events. The proposed scheme utilises a statistical method that goes through the whole time series from the start to the end, looking for the conditions that define events, according to the experts¿ criteria. Based on these extracted features, an Adaptive Fuzzy Inference Neural Network (AFINN) has been applied for the classification of stabilometric signals. The experimental results validated the proposed methodology.