GMixture
PURPOSE 
Gaussian mixture constructor.
SYNOPSIS
function gm=GMixture(varargin)
DESCRIPTION
Gaussian mixture constructor.
Defines a (non-necessarily normalized) Gaussian mixture given the
weights and the Gaussian for each component.
Possible parameters
- None: Defines an empty Gaussian mixture.
- Another Gaussian mixture: Copy constructor.
- w,g: The (row) vector of weigths and a cell array of Gaussians.
- g: The cell array of Gaussians. In this case the weights are
assumed uniform.CROSS-REFERENCE INFORMATION
This function calls:
This function is called by:
- GetTest1Parameters The example on Figure 1.
- MakeFigure4 Generates Figure 4.
- CompressGR Gaussian mixture compression using the Golberger and Roweis method.
- Crop Limits all means to be inside the given space.
- GMixture Gaussian mixture constructor.
- GetFirstComponents Gets the first components of a GMixture.
- GetLargestComponents Gets the components with the largest weigth.
- GetRandomComponents Selects components at random.
- Normalize Normalize a GMixture.
- Product gmOut=Product(gm1,gm2)
- RemoveSmallComponents Eliminates the small components of a GMixture.
- CS_CO_ObsModel CS_CO_ObsModel constructor.
- DiscretizeObsModel Discretizes the observation model on the observation side.
- GetObsModelFixedO Defines p(o|s) for a fixed 'o'.
- GetObsModelFixedS Defines p(o|s) for a fixed 's'.
- ComputeAlpha_a Compute the alpha_i_n-element for the given action and belief.
- ComputeAlpha_j_a_o Computes a particular alpha-element.
- DiscretizeRewardModel Discretizes the reward model on the action side.
- GetRewardModelFixedA Defines the reward function for a given action.
- UniformDistribution Uniform probability on a continuous space.
SOURCE CODE
0001 function gm=GMixture(varargin) 0002 % Gaussian mixture constructor. 0003 % 0004 % Defines a (non-necessarily normalized) Gaussian mixture given the 0005 % weights and the Gaussian for each component. 0006 % Possible parameters 0007 % - None: Defines an empty Gaussian mixture. 0008 % - Another Gaussian mixture: Copy constructor. 0009 % - w,g: The (row) vector of weigths and a cell array of Gaussians. 0010 % - g: The cell array of Gaussians. In this case the weights are 0011 % assumed uniform. 0012 0013 switch nargin 0014 case 0 0015 gm.w=[]; 0016 gm.g={}; 0017 gm.n=0; 0018 gm=class(gm,'GMixture'); 0019 case 1 0020 if isa(varargin{1},'GMixture') 0021 gm=varargin{1}; 0022 else 0023 if isa(varargin{1},'cell') 0024 n=size(varargin{1},2); 0025 gm=GMixture(ones(1,n)/n,varargin{1}); 0026 else 0027 error('Wrong type of parameters in GMixture constructor'); 0028 end 0029 end 0030 case 2 0031 n1=size(varargin{1},2); 0032 n2=size(varargin{2},2); 0033 if n1~=n2 0034 error('Size missmatch in GMixture constructor'); 0035 end 0036 if isa(varargin{1},'double') 0037 gm.w=varargin{1}; 0038 else 0039 error('Wrong type of parameters in GMixture constructor'); 0040 end 0041 if isa(varargin{2},'cell') 0042 gm.g=varargin{2}; 0043 else 0044 error('Wrong type of parameters in GMixture constructor'); 0045 end 0046 gm.n=n1; 0047 gm=class(gm,'GMixture'); 0048 otherwise 0049 error('Wrong number of parameters in GMixture constructor'); 0050 end