Statistical regularization
So, what is statistical regularization?
With regularization, whether we are speaking about mathematics, statistics, or machine learning, we are essentially talking about a process of adding additional information in order to solve a problem.
The term regularization has been described as an abstract concept of management of complex systems (according to a set of rules or accepted concepts). These rules will define how one can add or modify values in order to satisfy a requirement or solve a problem.
Does adding or modifying values mean changing data? (More about this will be studied later in this chapter.)
Various statistical regularization methods
Within the statistical community, the most popular statistical regularization methods may include the following:
- Ridge
- Lasso
- Least angles
Ridge
Ridge regression is a statistical technique that is used when analyzing regression data or models that suffer from a condition known as multicollinearity. When multicollinearity occurs...
