Efficient Intrusion Detection with KNN Classification and DS Theory - Proceedings of All India Seminar on Biomedical Engineering 2012 (AISOBE-2012) - page 166

Efficient Intrusion Detection with KNN
Classification and DS Theory
Deepika Dave and Sumit Vashishtha
Abstract Intrusion detection is an appallingly exigent area of research in the
existing scenario. Nowadays, to find a novel pattern of intrusions and detection is
an exceedingly difficult job. Our aim is to affect a method for intrusion detection
using KNN classification and Dempster theory of evidence. Using these modes, we
devised a new pattern of intrusion and classified category of pattern and applied
event evidence logic with the help of DS theory. Finned pattern of intrusion is
compared with the existing pattern of intrusion which generates a new schema of
pattern and updates a list of pattern of intrusion detection and improves the true
rate of intrusion detection. We have also accomplished some experimental tasks
with KDD99Cup and DARPA98 databases from MIT Lincoln Laboratory which
show that the proposed method provides competitively high detection rates
compared with other machine learning (ML) techniques and CRISP data mining.
The experimental results clearly show that the proposed system achieved higher
precision in identifying whether the records are abnormal or attacking ones.
Keywords Intrusion detection KNN DS theory KDD data set 99
D. Dave S. Vashishtha ( & )
Sagar Institute of Research, Technology and Science, Bhopal, Madhya Pradesh, INDIA
e-mail: [e-mail]
D. Dave
e-mail: [e-mail]