The Use of Ensemble Models for Multiple Class and Binary Class Classification for Improving Intrusion Detection Systems

. 2020 Apr 30;20(9):2559. doi: 10.3390/s20092559

Algorithm 3: Adaboost classifier algorithm

Input: KDD99 and nslkdd datasets
Training:

Selection of base classifier C;
Set the threshold for initial weights $W_{i 1} \in$ [0, 1], ${s u m}_{i = 1}^{N}$ $w_{i 1}$ = 1, Commonly $W_{i 1} = \frac{1}{N}$ ;
For n = 1 →k produce sample $D_{n}$ for training from D using the distribution $W_{n}$
Training of base classifier C on $D_{n}$ data subset to develop the $C_{n}$ classifier.
$e_{n} = \sum_{j = 1}^{N} w_{n i}$ is the ensemble error calculated when classifier $C_{n}$ misclassifies the $i^{t h}$ data point in D.
If $e_{n} \in$ (0, 0.5) then calculate $β_{n} = \frac{e_{n}}{1 - e_{n}}$ and update the next weight.
$w_{n + 1, i} = \{\begin{matrix} w_{n i} X β \\ w_{n i} \end{matrix}\}$
Distribution $W_{n + 1, i}$ needs to be normalized.
For further value of $e_{n}$ , set threshold $W_{i 1} = \frac{1}{N}$ and continue the process;
Return the trained classifiers $C_{1} C_{2} C_{3} . . . . . . . . . C_{n}$ and $β_{1} β_{2} β_{3} . . . . . . β_{n}$ .

Testing:

In the testing dataset, each data object X is passed to the testing dataset; classify by classifiers $C_{1} C_{2} C_{3} . . . . . . . . . C_{n}$ .
For each label y, assign to x by $C_{n}$ , calculate $m_{y} (x) = \sum C_{n} (x) = y l n (\frac{1}{β_{k}})$ . The class that has maximum value $m_{y} (x)$ is decided as the class label of x.
Repeat step2 for testing data and return the output.