CronoSeries.TimeSeries.Miscellaneous.LogLikelihoodPenalizer Class Reference

This class evaluates log-likelihood for a time series by summing up individual specified log-likelihoods. This is a rather trivial summation operation. What is useful here, however, is the consistency penalty. The idea is to look for a kind of draw-down in the partial sums of log-likelihoods and compute a penalty factor which is bad if the model is inconsistent with the data for a long sub-interval within the time range. More...

Public Member Functions
	LogLikelihoodPenalizer (Vector< double > components)
	The constructor takes as argument a Vector containing sequential individual Components of log-likelihood. More...

Properties
double	Penalty [get]
double	LogLikelihood [get]

Detailed Description

This class evaluates log-likelihood for a time series by summing up individual specified log-likelihoods. This is a rather trivial summation operation. What is useful here, however, is the consistency penalty. The idea is to look for a kind of draw-down in the partial sums of log-likelihoods and compute a penalty factor which is bad if the model is inconsistent with the data for a long sub-interval within the time range.

Constructor & Destructor Documentation

LogLikelihoodPenalizer()

CronoSeries.TimeSeries.Miscellaneous.LogLikelihoodPenalizer.LogLikelihoodPenalizer ( Vector< double >  components )

inline

The constructor takes as argument a Vector containing sequential individual Components of log-likelihood.

Parameters

v

---

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

• TimeSeries/Miscellaneous/LogLikelihoodPenalizer.cs