SISTEM DETEKSI INTRUSI JARINGAN DENGAN METODE SUPPORT VECTOR MACHINE
| Main Authors: | , AGUSTINUS JACOBUS, , Drs. Edi Winarko, M.Sc., Ph.D. |
|---|---|
| Format: | Thesis NonPeerReviewed |
| Terbitan: |
[Yogyakarta] : Universitas Gadjah Mada
, 2013
|
| Subjects: | |
| Online Access: |
https://repository.ugm.ac.id/118805/ http://etd.ugm.ac.id/index.php?mod=penelitian_detail&sub=PenelitianDetail&act=view&typ=html&buku_id=58781 |
Daftar Isi:
- Intrusion detection system is a system for detecting attacks or intrusions in a network or computer system, generally intrusion detection is done with comparing network traffic pattern with known attack pattern (misuse) or with finding unnormal pattern of network traffic (anomaly). The raise of internet activity has increase the number of packet data that must be analyzed for build the pattern that will be used for detect the intrusion, this situation led to the possibility that the system can not detect the attacks with a new technique or hidden. Using data mining methods are expected can solve this problem. This research having a goal to build an intrusion detection system in real-time environment by applying support vector machine method as a one of data mining method for classifying network traffic audit data (connection record) in 3 classes, namely: normal, probe, and DoS. Connection record was established from preprocessing of packet data header information that extracted from network packet capture files that obtained from network monitoring tools (tshark). Connection record is created in two forms, that is complete if the connection duration occurred less than 2 seconds and incomplete if the connection duration occurred more than 2 seconds. According to offline test using external data test, model from dataset DARPA KDD�99 give results accuracy 96,20%, attack detection rate 77,68%, and false positive rate 0,76%, and for model from simulation dataset obtained accuracy 90,14%, attack detection rate 66,60%, and false positive rate 3,50%. From the test in real-time condition, system successfully detect the intrusion action with accuracy 89,68%, attack detection rate 78,37%, and false positive rate 8,63%.