These are the most frequently used words in this book.
1994
activation
adaptive
al
algorithm
analysis
arbitrary
architectures
based
between
case
chapter
complex
conditions
convergence
data
derivative
dynamic
equation
equivalence
error
et
example
feedback
feedforward
figure
filter
first
fixed
form
function
gain
general
given
gradient
hence
however
input
iteration
layer
learning
linear
logistic
mandic
matrix
mean
model
modular
module
network
neural
neuron
nonlinear
nonlinearity
number
order
output
parameters
part
perceptron
performance
point
posteriori
prediction
predictor
problem
process
processing
rate
recurrent
referent
relationship
results
rnn
rtrl
series
set
shown
shows
sigmoid
signal
since
stability
state
static
structure
system
terms
therefore
time
trained
training
two
upon
use
used
values
vector
weight
whereas