Support Vector Machine (SVM) 3D

Question 1

A Support Vector Machine classifies by finding the…

Accepted Answer

The hyperplane maximizing the margin — SVM seeks the separating hyperplane with the largest gap (margin) to the nearest points of each class, improving generalization.

Question 2

The 'support vectors' are…

Accepted Answer

The points closest to the boundary — Only these boundary-defining points matter; moving other points doesn't change the hyperplane. This makes SVMs memory-efficient at prediction.

Question 3

Maximizing the MARGIN tends to improve…

Accepted Answer

Generalization, via the widest buffer — A wider margin is a form of regularization — it reduces sensitivity to individual points and lowers overfitting risk.

Question 4

The KERNEL TRICK lets SVMs…

Accepted Answer

Find boundaries via implicit mapping here — Kernels compute inner products in a high-dimensional feature space without ever computing the mapping explicitly, enabling non-linear separation efficiently.

Question 5

The RBF (Gaussian) kernel is popular because it…

Accepted Answer

Models complex, smooth non-linear boundaries — The RBF kernel measures similarity that decays with distance, allowing flexible curved boundaries; its gamma controls how local that influence is.

Question 6

The C hyperparameter in a soft-margin SVM controls…

Accepted Answer

The trade-off between margin width and errors — Large C penalizes misclassifications heavily (narrow margin, risk overfit); small C allows more violations for a wider, more robust margin.

Question 7

A 'soft margin' SVM is needed when…

Accepted Answer

Classes overlap — Real data is rarely perfectly separable; soft margins permit some points inside/over the margin, controlled by C.

Question 8

The loss function underlying the SVM is the…

Accepted Answer

Hinge loss — Hinge loss penalizes points within or on the wrong side of the margin, yielding the max-margin solution.

Question 9

The 'gamma' parameter in an RBF-kernel SVM controls…

Accepted Answer

How far one example's influence reaches — High gamma makes the boundary follow points tightly (risking overfit); low gamma gives smoother boundaries. Gamma and C are tuned together.

Question 10

Why is feature scaling important for SVMs?

Accepted Answer

Margins depend on distance, so scale distorts them — Like KNN, SVM distance/inner-product computations are dominated by large-range features unless standardized.

SVM Parameters

Data Parameters

Display

Fit a support vector machine

Test your knowledge

A Support Vector Machine classifies by finding the…

Interview Questions

References & Further Reading

Love this visualization?