Application of Partial-Connected Dynamic and GA-Optimized Neural Networks to Misuse Detection Using Categorized and Ranked Input Features

The  number  of  attacks  in  computer  networks  has  grown  extensively,  and many  new  intrusive methods  have  been appeared. Intrusion detection is known as an effective method to secure the information and communication systems. In  this paper,  the performance of Elman  and partial-connected dynamic neural network  (PCDNN)  architectures  are investigated for misuse detection in computer networks. To select the most significant features, logistic regression is also used to rank the input features of mentioned neural networks (NNs) based on the Chi-square values for different selected  subsets  in  this  work.  In  addition,  genetic  algorithm  (GA)  is  used  as  an  optimization  search  scheme  to determine  the  sub-optimal  architecture  of  investigated  NNs  with  selected  input  features.  International  knowledge discovery and data mining group  (KDD) dataset  is used  for  training and  test of  the mentioned models  in  this study. The  features  of  KDD  data  are  categorized  as  basic,  content,  time-based  traffic,  and  host-based  traffic  features. Empirical  results  show  that  PCDNN  with  selected  input  features  and  categorized  input  connections  offers  better detection  rate  (DR)  among  the  investigated models.  The mentioned NN  also  performs  better  in  terms  of  cost  per example (CPE) when compared to other proposed models in this study. False alarm rate (FAR) of the PCDNN with selected input features and categorized input connections is better than other proposed models, as well.

کلیدواژه ها:

en