Java Reference

In-Depth Information

produce an ouput, which may be subject to an
activation function
. An

activation function effectively is a transformation on the output,

which includes the specification of a
threshold
above which the out-

put is 1, otherwise zero. Figure 7-5(a) illustrates a neuron that takes

x
1,
x
2, and
x
3 input value and
w
1,
w
2, and
w
3 as input weights to

produce output value
y
.

Back
propagation
is the most common neural network learning

algorithm. It learns by iteratively processing the build data, comparing

the network's prediction for each case with the actual known target

value from the
validation data
.
3
For each case, the weights are updated

in the opposite direction, so as to minimize the error between the

network's prediction and actual target value.

Figure 7-5(b) illustrates a
back propagation neural network
that con-

sists of three types of layers:
input, hidden
, and
output
. The input

layer will have a number of neurons equal to the number of input

attributes, the output layer will have a number of neurons equal to

number of target values. The number of hidden layers and number

of neurons in each hidden layer can be determined by the algorithm

or specified by the data miner explicitly. In general, the addition of

a hidden layer can allow the network to learn more complex

patterns, but it can also adversely affect model build and apply

performance. For each neural layer, JDM allows specifying an

activation function
that computes the activation state of each neuron

N
(0,1)

W
(1,1,1)

W
(1,1,2)

x
1

N
(m-1,1)

N
(1,1)

N
(m,1)

y
1

•••

N
(0,2)

x
2

N
(m-1,2)

x
1

x
2

x
3

w
1

N
(1,2)

N
(m,2)

y
2

W
(1,3,2)

w
2

Neuron

y

N
(0,3)

x
3

w
3

Input Layer

Hidden Layer(s)

Output Layer

(a)

(b)

Figure 7-5

Neural networks: (a) Neuron representation, (b) back propagation neural

networks.

3

Validation data is a kind of test data used during model building, which the

algorithm may automatically create by partitioning the build data. Validation

data allows the algorithm to determine how well the model is learning the pat-

terns in the data. JDM allows users to provide an evaluation dataset explicitly in

a build task, if desired.

Search WWH ::

Custom Search