ML-Lec9-SVM

What is a Support Vector Machine (SVM)?

A powerful supervised machine learning algorithm used for classification, regression, and outlier detection tasks.

What is the primary goal of the SVM algorithm?

To create the best line or decision boundary that separates n-dimensional space into classes.

What is a hyperplane in the context of SVM?

The best decision boundary that creates the largest separation or margin between two classes.

What do support vectors represent?

The samples that are closest to the decision boundary and directly affect the location of the optimal decision boundary.

What are the two types of SVM?

Simple SVM, which is used for linear data, and Kernel SVM, which adds flexibility for non-linear data.

Why are SVMs considered effective for machine learning?

They can handle both classification and regression on linear and non-linear data, and they find complex relationships without extensive data transformations.

What are some applications of SVM?

Text classification, image classification, spam detection, handwriting identification, gene expression analysis, and face detection.

What is the significance of the maximum-margin hyperplane?

It represents the decision boundary that maximizes the margin between classes, leading to better generalization.

How does a linear SVM classifier work?

It makes a straight line between two classes, separating the data points into categories.

What is the primary disadvantage of using SVMs for large datasets?

SVMs work best on small sample sets due to high training time.

What is the distance expression for a point x to a line wx + b = 0?

The distance can be computed using the expression that involves the weight vector and normal vector.

How can SVM ignore outliers?

By finding the best hyperplane that maximizes the margin while maintaining robustness to outliers.

What is a kernel in SVM?

A function that transforms the data into a higher dimension to allow for a more flexible decision boundary.

What characterizes the decision boundary created by SVMs?

It can be non-linear and can separate data points into classes through various kernel functions.

What is a linear kernel?

A commonly used kernel for text classification that is faster and suited for linearly separable data.

What does the term 'margin' refer to in SVMs?

The distance between the hyperplane and the closest data points from either class.