This
above mentioned algorithm generates the dictionary that can be used to extract
the feature vectors for the long payload features. Figure 2 shows the block diagram
of dictionary construction. Now this dictionary which has been built with
Bigrams and Trigrams can be used by any feature extraction algorithm more
effectively even from the long payload features. The generated Bigrams and
Trigrams can also be handled by any machine learning algorithms. In the present
work, experiments were carried with one of the most widely cited machine
learning algorithm, Support Vector Machine (SVM) to study the results out of
the proposed methodology.  

The
further step is the feature vector extraction for the long payload features
which is outlined in the following figure 3. The approach in 16 has been
followed in the present work for the task of feature vector extraction. The
feature vector extraction step is also explained in more detail in the below
Algorithm 2.

______________________________________

Algorithm 2:
Feature Vector Extraction for Long Payload Features with Bigrams & Trigrams

Step-1: Input
the long payload features and   

           constructed Bigram / Trigram
Dictionary

Step-2:
Initialize all feature vectors to zero

Step-3: Take one
string at a time 

Step-4: Take one
Bigram / Trigram 

Step-5: Find the
index of Bigram / Trigram

           from the dictionary

Step-6:
Increment the location counter in the

           feature vector

Step-7: Repeat
till there is no possibility of

           finding Bigram / Trigram from the
input

           payload feature

Step-8: Finish
feature vector i, and proceed to

           feature vector i+1

Step-9: Stop the
process if all there no feature

            vector is left to be processed

Step-10: Output
the feature vectors