This code is based on the paper:
"Inferring social network structure in ecological systems from spatio-temporal data streams"
by Ioannis Psorakis, Stephen J. Roberts, Iead Rezek and Ben C. Sheldon,
Journal of the Royal Society Interface, vol. 9 no. 76, pp 3055-3066 (2012).

The following colleagues have also provided invaluable contributions, insights, bug reports 
and feature suggestions:

Damien Farine
Bernhard Voelkl
Colin Garroway
Reinder Radesma
Lucy Aplin

along with the rest of the Edward Grey Institute team at Oxford University Dept. of Zoology.

------
The package contains two scripts, run from the MATLAB command line:
gmmevents.m: extracts gathering events and builds social network of statistically significant links
gmmevents_groups_only.m: extracts ONLY the gathering events from the data stream

The two scripts have different outputs and for more information please write on MATLAB prompt:
>> help gmmevents
%or
>> help gmmevents_groups_only
or see file description in the comments.

Both scripts have a common input structure, which is:
% INPUTS:
% -------
% DATA: a Z-by-3 matrix, where Z the total number of observations. The
% first column Z(:,1) denotes the real-valued timestamps of each observation. The
% second column Z(:,2) denotes the indices of the individuals that were
% recorded. Such indices must be INTEGER values, as they refer to nodes in
% the adjacency matrix. Finally, the third column Z(:,3) denotes the index
% of the location where the observation took place. Again, such indices
% must be integer values.
% Example:
% % if we wish to see the contents of the 15th observation (z=15) in our
% data stream, we write:
% >> DATA(15,:)
% ans =
%    18516495           2           1
%
% which reads: "the individual i=2 appeared at location l=1 at time
% t=18516495"
%
% N: is the total number of distinct individuals in the data stream. If not
% provided, it is automatically calculated by the second column of DATA

% number_of_randomisations: the total number of data stream permutations
% during link significance testing. If the user does not wish to perform
% the test, set the value of that variable less or equal than 1, or not
% pass it at all.

--------