Optimisation of machine learning algorithm
Objective: find the parameters of the K-nearest neighbors classifier that maximise the cross-validation score and the accuracy
Optimisation algorithm: Indicator-Based Evolutionary Algorithm (IBEA)
Parameters to optimise: algorithm, leaf_size, n_neighbors, p, weights (Please refer to sklearn.neighbors.KNeighborsClassifier documentation). The decision vector is coded into binary.
Objective: minimise the negative of the cross-validation score and the accuracy score.
To run the optimisation (Matlab): [Archive,Archive_objectives, X, Xo, samples, samples_objectives] = IBEA_binary(10, 5, 'cost_func', 30, 2, 0.1, 1, 0.01, 0.05)
To save the results: save IBEA.mat
To visualise the results: plot_results
To write the best solution to a file: write_best(index_best)