Algorithm 2. Proposed HNIDS method (IRF working)

Input: NSL dataset ds for training, v variable, n1 represents the total the nodes in a tree Output: Random Forest ensemble tree RF$\mathrm{Tree}(\begin{array}{c}\mathrm{k}\\ 1\end{array})$

Step1: Construct an initial tree

1.1 for each j to k repeat

1.2 construct a sample set by using an initial IDS dataset (original) of size n

1.3 start feeding the new bootstrapped dataset to a decision tree DFTree

Step2: Choose the best fit

2.1 for each n1 to min node size, repeat

2.2 start feeding the new bootstrapped dataset to a Random forest tree RFTree

2.3 Select randomly variables v’, towards a variable set v

2.4 Choose the fittest variable divided among all these variables v’

2.5 divide a parent tree node into the new child nodes

2.6 return ensemble tree RF$\mathrm{Tree}(\begin{array}{c}\mathrm{k}\\ 1\end{array})$

2.7 End

Step 3: Verify a constructed Random Forest ensemble tree RF$\mathrm{Tree}(\begin{array}{c}\mathrm{k}\\ 1\end{array})$

3.1 if (ensemble tree RF$\mathrm{Tree}(\begin{array}{c}\mathrm{k}\\ 1\end{array})$ == Best_fit_tree BF$\mathrm{Tree}(\begin{array}{c}\mathrm{k}\\ 1\end{array}))$

3.2 Proceed with the best fir ensemble tree

Step 4: Determine the best class

4.1 apply classification to find the best fit from BF$\mathrm{Tree}(\begin{array}{c}\mathrm{k}\\ 1\end{array})$

4.2 Calculate the number of votes

4.3 if the number of votes is maximum

4.4 return BF$\mathrm{Tree}(\begin{array}{c}\mathrm{k}\\ 1\end{array})$

4.5 End