The Development and Evaluation of a Dataset for Testing of IDS for Wireless Networks

This paper originates from research that investigated about the creation of the dataset representative of a wireless computer network. The proposal was intended to generate the dataset from network traffic of a real wireless network to be employed in the evaluation of Intrusion Detection Systems - IDS. Several attacks were granted against the network in order to obtain data from these anomalous behaviors. The methodological procedures performed involved the capture of the traffic on the wireless network and it's pre-processing to generate the dataset. To evaluate the dataset the following pattern classification algorithms were employed: Bayesian Networks, Decision Tables, IBK, J48, MLP and NaiveBayes, which are generally used in implementation of IDS. In addition, the Kappa coefficient was also used to assist in measure of the efficiency of algorithms employed. The good results obtained show that the data set can be used to compare different approaches of IDS for wireless networking.