ZeroMean<T>

Models & Types · AiDotNet.GaussianProcesses

Implements a zero mean function.

Implements: IMeanFunction<T>

For Beginners

The zero mean function returns zero for all inputs. This is the most common default choice for GPs.

m(x) = 0

Why use zero mean?

  • Simplicity: No parameters to tune
  • Flexibility: The GP can still learn any function through the kernel
  • Data centering: Often we center our data to have zero mean anyway

When zero mean makes sense:

  • Centered/normalized data
  • When you want the GP to be flexible
  • When you don’t have strong prior beliefs about the trend

Constructors

Constructor Summary
ZeroMean() Initializes a new zero mean function.

Methods

Method Summary
T Evaluate(Vector<T> x) Returns a vector of zeros for all input points.
Vector<T> Evaluate(Matrix<T> X) Returns a vector of zeros for all input points.