Nowadays, Security of network traffic is becoming a major issue of
computer network system according to the huge development of internet.
Intrusion detection system has been used for discovering intrusion and to
maintain the security information from attacks. In this paper, produced two
levels of mining algorithms to construct Network Intrusion Detection System
(NIDS) and to reduce false alarm rate, in the first level Naïve Bayes algorithm
is used to classify abnormal activity into the main four attack types from
normal behavior. In the second level ID3 decision tree algorithm is used to
classify four attack types into (22) children of attacks from normal behavior.
To evaluate the performance of the two proposed algorithms by using kdd99
dataset intrusion detection system and the evaluation metric accuracy,
precision, DR, F-measure. The experimental results prove that the proposal
system done high detection rates (DR) of 99 % and reduce false positives (FP)
of 0 % for different types of network intrusions.