Kernel functions are used in machine learning including support vector machines to increase non-linear diversity. The term originates in mathematics where a postive definite kernel is a form of non-linear generalisation of the inner prooduct.
Used in Chap. 7: page 99; Chap. 8: page 104
Links:
Wikipedia: Kernel method- Wikipedia: Positive definite kernel
- turing.com: Designing of Different Kernels in Machine Learning and Deep Learning