Hands-On Artificial Intelligence for Cybersecurity

Making decisions with trees

As we saw in the previous paragraphs, when we have to choose which algorithm to use to perform a given task, we must consider the type of features that characterize our data. The features can in fact be made up of quantitative values ​​or qualitative data.

ML algorithms are obviously more at ease when dealing with quantitative values; however, most of the real cases involve the use of data expressed in a qualitative form (such as descriptions, labels, words, and so on) that imply information expressed in non-numerical form.

As in the case of spam detection, we have seen how the translation in numerical form (a practice known as numeric encoding) of qualitative features (such as the spam and ham labels, to which ...