Data Input Argument to Functions

In MixEst, all the functions taking a data input argument, can accept the data in one of the following forms:

Data Matrix

In matrix form, the data is arranged in an n-by-N matrix where n is the dimensions of the data space and N is the number of data points (i.e. each column of the matrix represents a single observation).

Example

% create a distribution and random parameters
D = mvnfactory(2);
theta = D.randparam();
% generate 1000 random 2-dimensional data points
data = randn(2, 1000);
% pass the data to the ll function
ll = D.ll(theta, data)

Note: In MATLAB versions R2011b and above, instead of the data matrix, you can give a matlab.io.MatFile object referring to a MAT-file containing a variable named data with specifications as pointed out above.

Data Structure

Using the structure form, you can pass along with the data matrix, the indices of a subset of data points and/or a weighting vector on the data points whenever applicable. In this form, data is a structure with the following fields (All fields except data are optional):

  • data (n-by-N) The data matrix as described in the previous section.
  • index (1-by-L) Data indexing vector for selecting a subset of data (Only numeric indexing is supported. Logical indexing is not supported).
  • weight (1-by-N, or 1-by-L if indexing is used) Data weighting vector containing values between zero and one to be multiplied by each data point.

where n is the dimensions of the data space, N is the total number of data points, and L is the number of points in the indexed subset of data.

Example 1

% create a distribution and random parameters
D = mvnfactory(2);
theta = D.randparam();
% generate 1000 random 2-dimensional data points
data = randn(2, 1000);
% use only the 100th up to 200th points from the data
index = 100:200;
% generate random weights for the selected data
weight = rand(1, 101); % 101 is the number of elements in index
% build the data structure
data = struct('data', data, 'index', index, 'weight', weight);
% pass the data to the ll function
ll = D.ll(theta, data)

Example 2

D = mvnfactory(2);
theta = D.randparam();
ll = D.ll(theta, struct('data', randn(2,1000), 'weight', rand(1,1000)))