ARTIFICIAL NEURAL NETWORK the net input can be calculated as

follows
IT an efficient computing system
whose central theme is borrowed
yin=x1.w1+x2.w2+x3.w3…xm.wmy
from the analogy of biological neural
in=x1.w1+x2.w2+x3.w3…xm.wm
networks. ANNs are also named as i.e., Net input yin=∑mixi.wi
“artificial neural systems,” or “parallel output can be calculated by applying
distributed processing systems,” or the activation function over the net
“connectionist systems.” input.
ANN acquires a large collection of units Y=F(yin) Output = function
that are interconnected in some pattern
to allow communication between the
Radial basis function (RBF) networks
units. These units, also referred to as
are a commonly used type of
nodes or neurons, are simple
artificial neural network for function
processors which operate in parallel.
approximation problems. Radial
Every neuron is connected with other basis function networks are
neuron through a connection link. Each distinguished from other neural
connection link is associated with a networks due to their universal
weight that has information about the approximation and faster learning
input signal. This is the most useful speed.
information for neurons to solve a
particular problem because the weight
usually excites or inhibits the signal that
is being communicated. Each neuron
has an internal state, which is called an
activation signal. Output signals, which
are produced after combining the input
signals and activation rule, may be sent
to other units.
Model of Artificial Neural Network
Functional link artificial neural
network (FLANN) is a single layer
ANN with low computational
complexity which has been used
in versatile fields of application,
such as system identification,
pattern recognition, prediction
and classification, etc
SOMA– NODE
DENDRITES – INPUT
SYNAPSE – WEIGHTS
AXON - OUTPUT
Backpropagation defines the whole distribute the error at the output unit
process encompassing both the Yk back to all units in the previous layer.
calculation of the gradient and its need
Similarly the factor δδj(j = 1, ... . p) is
in the stochastic gradient descent.
compared for each hidden unit Z j.
Technically, backpropagation is used to
calculate the gradient of the error of the It can update the weights and
network concerning the network’s
biases. Types of Backpropagation:-
modifiable weights. The characteristics
of Backpropagation are the iterative, Static Back Propagation − In this type
recursive and effective approach of backpropagation, the static output is
through which it computes the updated created because of the mapping of
weight to increase the network until it is static input. It is used to resolve static
not able to implement the service for
classification problems like optical
which it is being trained.
Backpropagation is widely used in character recognition. Recurrent
neural network training and calculates Backpropagation − The Recurrent
the loss function for the weights of the Propagation is directed forward or
network. Its service with a multi-layer directed until a specific determined
neural network and discover the
internal description of input-output value or threshold value is acquired.
mapping. It is a standard form of After the certain value, the error is
artificial network training, which evaluated and propagated backward.
supports computing gradient loss
function concerning all weights in the
network. The backpropagation
algorithm is used to train a neural
network more effectively through a
chain rule method.
The training algorithm of
backpropagation involves four stages :-
Initialization of weights − There are
some small random values are
assigned. Feed-forward − Each unit X
receives an input signal and transmits
this signal to each of the hidden unit Z 1,
Z2,... Zn. Each hidden unit calculates the
activation function and sends its signal
Z1 to each output unit. The output unit
calculates the activation function to
form the response of the given input
pattern. Backpropagation of errors −
Each output unit compares activation
Yk with the target value Tk to determine
the associated error for that unit. It is
based on the error, the factor δδk(K = 1,
... . m) is computed and is used to