Efficient Intrusion Detection with KNN Classification and DS Theory - Proceedings of All India Seminar on Biomedical Engineering 2012 (AISOBE-2012) - page 177

Fig. 4 The overall steps of the proposed intrusion detection system

and testing data set. Initially, the training data set is classified into five subsets so
that, four types of attacks (DoS, R2L, U2R, Probe) and normal data are separated.
After that, we simply mine the one-length frequent items from attack data as well
as normal data. These mined frequent items are used to find the important
attributes of the input data set and the identified effective attributes are used to
generate a set of definite and indefinite rules using deviation method. Then, we
generate fuzzy rule in accordance with the definite rule by fuzzifying it in such a
way that we obtain a set of fuzzy if-then rules with consequent parts that rep-
resent whether it is a normal data or an abnormal data. These rules are given to
the fuzzy rule base to effectively learn the fuzzy system. In the testing phase, the
test data is matched with fuzzy rules to detect whether the test data is an
abnormal data or a normal data. We apply KNN classification and Dempster
theory of evidence on classified data. Through these we gathered a new
discovered pattern of intrusion and classified category of pattern and apply event
evidence logic with the help of DS theory. Finned pattern of intrusion compared
with the existing pattern of intrusion generates a new schema of pattern and
updates a list of patterns of intrusion detection and improved the true rate of
intrusion detection. We used the concept of Dempster theory for this work on
event evidence and found the validity of data and reduced the rate of intrusion.
We also used the patterns of design of schema and data conversion, in data
conversion first-type intrusion detection in MATLAB, but data of intrusion data
in overall string format, now we has use classification method. We have faced
various difficulties in classification of data conversion string through numeric
format for suitability of classification. In the process of data conversion we used
the ratio mapping concept used by the ML receptory organization for mapping of
data string to numeric format. The above procedure is explained and depicted
with the help of a flowchart in (Fig. 4 );