Projection pursuit strategies. More...

Classes
struct	PPPDAStrategy
	Penalized Discriminant Analysis projection pursuit strategy. More...

struct	PPResult
	Result of a projection pursuit optimization step. More...

struct	PPStrategy
	Abstract strategy for projection pursuit optimization. More...

Typedefs
using	Projector = types::FeatureVector
	Column vector of projection coefficients (one per variable).

Functions
Projector	normalize (Projector const &projector)
	Normalize a projector to unit length.

PPStrategy::Ptr	pda (float lambda)
	Factory function for a PDA projection pursuit strategy.

Detailed Description

Projection pursuit strategies.

Contains the abstract PPStrategy interface and concrete implementations (e.g. PPPDAStrategy) that define how to evaluate and optimise a projection index for separating groups.

New strategies must implement the pure virtual methods including to_json() for serialization support.

Typedef Documentation

◆ Projector

using ppforest2::pp::Projector = types::FeatureVector

Column vector of projection coefficients (one per variable).

Function Documentation

◆ normalize()

Projector ppforest2::pp::normalize ( Projector const & projector )

Normalize a projector to unit length.

Parameters

projector Input projector (p).

Returns: Normalized projector with L2 norm = 1.

◆ pda()

PPStrategy::Ptr ppforest2::pp::pda ( float lambda )

Factory function for a PDA projection pursuit strategy.

Classes