Quiz2

Support Vector Machines are primarily designed for:

• Regression problems to predict continuous values.

• Unsupervised learning tasks like clustering data points.

• Classification problems to separate data points into distinct classes.

• Both regression and classification tasks equally well.

Which of the following is NOT an advantage of using SVMs?

• Effective in high-dimensional spaces due to the 'curse of dimensionality' being less impactful.

• Efficient for prediction once the model is trained.

• Requires careful selection and tuning of hyperparameters.

• Generally performs well with small datasets.

In the context of SVMs, what are 'support vectors'?

• Randomly chosen data points used to define the hyperplane.

• Data points closest to the hyperplane that define the margin.

• Data points that are misclassified by the model.

• All data points used to train the SVM.

A high value of the regularization parameter (C) in an SVM indicates:

• A more complex model with a smaller margin but potentially better handling of outliers.

• A simpler model with a larger margin but stricter penalty for misclassification.

• There is no impact on the model complexity.

• The model will always overfit the training data.

The primary objective of a Support Vector Machine (SVM) is to:

• Minimize the distance between data points and the hyperplane.

• Maximize the variance of the data points.

• Find the hyperplane with the largest number of data points on it.

• Minimize the error on the training data, regardless of margins.

What type of kernel function is typically used in SVMs for non-linearly separable data?

• Linear kernel

• Polynomial kernel

• Sigmoid kernel

• All of the above (depending on the scenario)

SVMs are generally considered to be:

• Highly interpretable due to the clear decision boundary.

• Prone to overfitting, especially with high-dimensional data.

• Not suitable for multi-class classification problems.

• All of the above.