Information Technology Reference
In-Depth Information
4 Experiment and Analysis of Intrusion Detection System
Based on MLP Algorithm
MLP is conceivably the most popular network architecture currently in use amongst
the ANNs (Saftoiu et al.
2012
). There are three layers of units: input layer, a hidden
layer and an output layer in the architecture of MLP with feed-forward supervised
learning. The proposed ANN architecture was implemented using the SPSS neural
networks program using SPSS 16.0 (
http://www-01.ibm.com/software/in/analytics/
spss/downloads.html
)
in Windows XP environment. Neural Networks are nonlinear
statistical data modeling approaches. ANNs can explore and extract nonlinear
interactions among parameters to expose formerly unidenti
ed associations among
given input parameters and outcomes (Sall et al.
2007
).
The Fig.
8
shows a feed forward architecture of the neural network because the
connections in the network
flow forward from the input layer to the output layer
without any feedback loops. In this Fig.
8
the input layer contains the 39 predictors;
one hidden layer contains unobservable nodes, or units. Based on some function of
fl
Fig. 8 Feedforward architecture with one hidden layer
Search WWH ::
Custom Search