Hands-On Artificial Intelligence for Cybersecurity

SVM optimization strategy

Why choose the hyperplane that maximizes the margin in the first place? The reason lies in the fact that wider margins correspond to fewer classification errors, while with narrower margins we risk incurring the phenomenon known as overfitting (a real disaster that we may incur when dealing with iterative algorithms, as we will see when we will discuss verification and optimization strategies for our AI solutions).

We can translate the SVM optimization strategy in mathematical terms, similar to what we have done in the case of the Perceptron (which remains our starting point). We define the condition that must be met to assure that the SVM correctly identifies the best hyperplane that separates the classes of data: ...