/numerical-optimization

Primary LanguagePython

Reference Codebook

Numerical Optimization Routines from Scratch

Contents:

Base Classes:

  • ProbabilityDistribution: base template class for all probability distributions
  • Optimizer: base template class for performing single-output function optimization
  • Kernel: base template class for kernel transformations

Probability Distribution Classes:

  • NormalDistribution1D(ProbabilityDistribution): Gaussian/normal distribution for one dimension
  • NormalDistribution2D(ProbabilityDistribution): Gaussian/normal bivariate distribution
  • MVNormalDistribution(ProbabilityDistribution): Multivariate Gaussian/normal distribution

Kernel Classes:

  • LinearKernel(Kernel): Linear (dot product) kernel
  • RBFKernel(Kernel): Gaussian/radial basis function kernel (AKA squared exponential kernel)
  • RationalQuadraticKernel(Kernel): Rational quadratic kernel

Optimizer Classes:

  • TBD