Statistics::LogisticModel Class Reference

Models the probability of a boolean outcome using a logit. More...

#include <LogisticModel.h>

Public Member Functions
Real	Evaluate (const Vector &x) const
Real	EvaluateExpectation (const Vector &x, int numSamples=20) const
Real	EvaluateVariance (const Vector &x, int numSamples=20) const
void	MaximumLikelihood (const std::vector< Vector > &x, const std::vector< bool > &v, int numIters, Real tol)

Public Attributes
Vector	coeffs
Matrix	covariance

Detailed Description

Models the probability of a boolean outcome using a logit.

The function gives p(x)=e^v/(1+e^v), where v = coeffs^t*x The data must contain a constant term in position n. This requirement may be relaxed in future versions

---

The documentation for this class was generated from the following files:

• LogisticModel.h
• LogisticModel.cpp