By Sam Berens (s.berens@sussex.ac.uk)

Click here to download release R02.00



Highlights of this release

HoopStats_EstimMixtureModel.m
Estimates a mixture model for circularly distributed data. Components within the model can include a uniform distribution and arbitrarily many target/fixed position von Mises distributions.

HoopStats_RunEM.m
Estimates a circular mixture model using Expectation-Maximization (EM).

HoopStats_HardCluster.m
Estimates a mixture model for circularly distributed accuracy data using a hard clustering method. At present, this implementation is only setup to fit one target distribution and one uniform component.

HoopStats_CalcParams.m
Computes various statistics for a sample of angles in radians ("Thetas"). If optional "Weights" argument is supplied, angles are weighted to have varying influences on the statistics.

HoopStats_InfoContent.m
Computes the information content (I) of a von Mises distribution with a prior weight of P and a concentration parameter of K.

HoopStats_K2R.m
Converts the von Mises concentration parameter (K) to a resultant (mean) vector length R.

HoopStats_KsDensity.m
Returns kernel density estimates for a sample of circularly distributed data in "P", evaluated at the angles in "Theta". "K" is a kappa value for the von Mises distribution. Here, it defines the kernel bandwidth of the that spreads the density function around each point in "P".