Source code for bcselector.filter_methods.no_cost_based_filter_methods

import numpy as np

[docs]def no_cost_find_best_feature(j_criterion_func, data, target_variable, possible_variables_index, costs, **kwargs): variables_result = [] for i in possible_variables_index: variables_result.append(j_criterion_func(data, target_variable = target_variable, candidate_variable_index=i, **kwargs)) k = np.argmax(variables_result) return possible_variables_index[k], variables_result[k], costs[possible_variables_index[k]]