Contents
function [P_fp_est,P_fn_est,Stat] = estimate_ROC(P_fp_target,dim,varargin)
%[P_fp_est,P_fn_est,Stat] = estimate_ROC(P_fp_target,dim,option) % % Inputs % * dim: Dimension of the particule space % * P_fp_target: Required false positive probabilities % * option: a structure containing the following parameters (fields) % % Parameters % * H0_setup: a structure defining the setup for H0 simulation % compliant with the estimator_B option structure % * H1_setup: a structure defining the setup for H1 simulation % compliant with the estimator_B option structure % % Outputs % * P_fp_est: estimated false positivie probabilities % * P_fn_est: estimated false positivie probabilities % * Stat : a structure containing some statistics about the estimation % % Comments % Authors: % * Implementation: Teddy Furon % * Science: Frederic Cerou, Pierre Del Moral, Arnaud Guyader, Teddy Furon % % Contact: teddy.furon@inria.fr % % Copyright INRIA 2010 % % Licence: % This software and its subprograms are governed by the CeCILL license % under French law and abiding by the rules of distribution of free software. % See http://www.cecill.info/licences.en.html
Parsing the arguments
error(nargoutchk(0,3,nargout)); % Default setups H0_default.verbose = false; H0_default.n = 150; H0_default.cell = false; H1_default.verbose = false; H1_default.n = 300; H1_default.cell = false; sigw = 0.1; % watermark power H1_default.SCORE = @(x)SCORE_H1(x,sigw); % Create an instance of the inputParser class. p = inputParser; % Define inputs that one must pass on every call: p.addRequired('dim', @(x)validateattributes(x,{'numeric'},... {'integer', 'positive','scalar'})); p.addRequired('P_fp_target',@(x)validateattributes(x,{'numeric'},... {'real', 'nonempty','positive'})); p.addParamValue('H1_setup',H1_default,@(x)isa(x,'struct')); p.addParamValue('H0_setup',H0_default,@(x)isa(x,'struct')); % Parse and validate all input arguments. p.StructExpand = true; p.KeepUnmatched = true; p.parse(dim,P_fp_target,varargin{:}); h = p.Results;
Input argument "dim" is undefined.
Error in ==> estimate_ROC at 62
p.parse(dim,P_fp_target,varargin{:});
Hypothesis H0
Estimate thresholds for some false positive probabilties
if ~isfield(h.H0_setup,'cell') h.H0_setup.cell=false; end if h.H0_setup.cell==true % Are you working with cells? % Warm up: For required P_fp, we first estimate the associated thresholds q= estimator_quantile_B_cell(h.P_fp_target,h.dim,h.H0_setup); % Now, for this thresholds, we estimate P_fp [P_fp_est,Stat_fp,Intern_fp] = estimator_B_cell(q,h.dim,h.H0_setup); else % Warm up: For required P_fp, we first estimate the associated thresholds q= estimator_quantile_B(h.P_fp_target,h.dim,h.H0_setup); % Now, for this thresholds, we estimate P_fp [P_fp_est,Stat_fp,Intern_fp] = estimator_B(q,h.dim,h.H0_setup); end
Hypothesis H1
Estimate probabilities for some thresholds
if ~isfield(h.H1_setup,'cell') h.H1_setup=false; end if h.H1_setup.cell==true % Are you working with cells? [P_fn_est,Stat_fn,Intern_fn] = estimator_B_cell(-q,dim,h.H1_setup); else [P_fn_est,Stat_fn,Intern_fn] = estimator_B(-q,dim,h.H1_setup); end
Build Stat
Stat.interval_fp = [Stat_fp.interval]; Stat.interval_fn = [Stat_fn.interval]; Stat.threshold = q;
Subfunctions
Default functions
function y = SCORE_H1(X,sigw)
X(1,:) = X(1,:) + sigw*sqrt(size(X,1));
y = - abs(X(1,:))./sqrt(sum(X.^2,1));