[Fund. of Statistical Learning] Nomparametric Methods

[cheatsheet1999]

Basic Idea of Density Estimation

• "Density" p(x) describes how densely the samples are likely to appear around x
  • Basic idea in density estimation: the more data in a region, the larger is the density function in that region
  • Prob that a vector x will fall in region R is

• Given some training samples, the left-hand side can be estimated as

• V is the volume of integration