Hybrid Correlation-Neural Network Synergy for Gait Signal Classification - Advances in Heuristic Signal Processing and Applications - page 266

Fig. 12.5 Magnitude cross-spectral density of the left stance interval belonging to healthy and
various pathological subjects
Fig. 12.6 Phase cross-spectral density of the left stance interval belonging to healthy and various
pathological subjects
only feedforward connection, RNNs are more complicated as they employ a combi-
nation of feedback and feedforward connections, exhibiting the property of memory
[ 4 , 18 ]. RNNs are useful for two types of application: as associated memories and
for input-output mapping [ 12 , 40 , 44 , 48 ]. In the present application, RNNs have
been implemented for input-output mapping. There are several architecture layouts
available for different relevant RNNs. Some popular variants of such RNNs include
Jordan's network (which employs feedback connection from the output of the output