[bn_fit_corr] = BN_Fitness_Corr(bnet,Pop,FunVal): BN_Fitness_Corr: Computes the correlation between the probabilities given to the solutions by the networks and the fitness values of the solutions INPUTS bnet: Bayesian network Pop: Population FunVal: Evaluation vector for each of the objectives OUTPUTS bn_fit_corr: Vector of correlations for each of the objectives. Last version 8/26/2008. Roberto Santana (roberto.santana@ehu.es)
0001 function[bn_fit_corr] = BN_Fitness_Corr(bnet,Pop,FunVal) 0002 % [bn_fit_corr] = BN_Fitness_Corr(bnet,Pop,FunVal): 0003 % BN_Fitness_Corr: Computes the correlation between the probabilities given to the solutions by the 0004 % networks and the fitness values of the solutions 0005 % INPUTS 0006 % bnet: Bayesian network 0007 % Pop: Population 0008 % FunVal: Evaluation vector for each of the objectives 0009 % OUTPUTS 0010 % bn_fit_corr: Vector of correlations for each of the objectives. 0011 % 0012 % Last version 8/26/2008. Roberto Santana (roberto.santana@ehu.es) 0013 0014 0015 [Prob] = BN_Pop_Prob(bnet,Pop); 0016 0017 number_objectives = size(FunVal,2); 0018 0019 for i=1:number_objectives, 0020 bn_fit_corr(i) = corr(Prob,FunVal(:,i)); 0021 end, 0022 0023 0024 0025