November 2019
Intermediate to advanced
346 pages
9h 36m
English
We begin Step 1 by loading the data by unpickling it. The dataset has been pre-engineered to be balanced, so we do not need to worry about imbalanced data challenges. In practice, the detection of botnets may require satisfying a constraint on false positives. Moving on, we utilize the already predefined train-test split to split our data (Step 2). We can now instantiate a classifier, fit it to the data, and then test it (Steps 3 and 5). Looking at the accuracy, we see that it is quite high. Since the dataset is already balanced, we need not worry that our metric is misleading. In general, detecting botnets can be challenging. The difficulty in detecting botnets is illustrated by the GameOver Zeus botnet malware package. Originally ...