Definition of a metric for medical symptoms using neural networks
R. Silva, A.C.R. da Silva · 2003
This work describes a new way of relating medical symptoms among themselves via the definition of a distance measure. Medical symptoms are represented by sinusoidal damped waves whose parameters are obtained via a semantic analysis. These wave patterns are discretized and presented to a self-organizing neural network previously trained with a single signal. The outputs provide a numerical classification for the symptoms and define a metric relationship among them.