Algorithm 2 Genetic

Input: X, y, X_train, X_test, y_train, y_test, new_model Output: Gene, score Begin new_model = ML Function initilization_of_population(size, n_feat)         population = Empty list         For i in range from 1 to size                 chromosome = Boolean array of size n_feat initialized to True                 For j in range from 0 to int(0.3 * n_feat)                         chromosome[j] = False                 Endfor                 Randomly shuffle the elements in chromosome                 Append chromosome to the population         Endfor       Return population Endfunction Function fitness_score (population, X_train, X_test, y_train, y_test)         scores = Empty list         For each chromosome in population                 Create a model (new_model)                 Train the model on X_train using the features specified by the chromosome                 Predict class labels on X_test with the model                 Calculate accuracy score by comparing predictions with y_test                 Append the score to scores           Endfor         Sort scores in descending order while maintaining the correspondence with the population         Return scores, sorted population Endfunction Function selection (pop_after_fit, n_parents)         population_nextgen = Empty list         For i in range from 1 to n_parents                 Append pop_after_fit[i] to population_nextgen         Endfor         Return population_nextgen Endfunction Function crossover (pop_after_sel)         pop_nextgen = Copy pop_after_sel         For i in range from 0 to the size of pop_after_sel - 1 with a step of 2                 new_par = Empty list                 child_1, child_2 = pop_nextgen[i], pop_nextgen[i+1]                 new_par = Concatenate the first half of child_1 and the second half of child_2.                 Append new_par to pop_nextgen               Endfor         Return pop_nextgen Endfunction Function mutation (pop_after_cross, mutation_rate, n_feat)         mutation_range = Integer rounded down of mutation_rate times n_feat         pop_next_gen = Empty list         For each individual in pop_after_cross                 chromosome < = Copy the individual                 rand_posi < = Empty list               For i in range from 1 to mutation_range                         pos < = Random integer between 0 and n_feat - 1                         Append pos to rand_posi                 Endfor                 For each position in rand_posi                         Invert the value of the gene corresponding to that position in chromosome                 Endfor               Append chromosome to pop_next_gen         Endfor       Return pop_next_gen Endfunction Function generations (df, label, size, n_feat, n_parents, mutation_rate, n_gen, X_train, X_test, y_train, y_test)         best_chromo = Empty list         best_score = Empty list         population_nextgen = Call initilization_of_population with size and n_feat         For each generation i from 1 to n_gen                 scores, pop_after_fit = Call fitness_score with population_nextgen, X_train, X_test, y_train, y_test                 Print the best score in generation i                 pop_after_sel = Call selection with pop_after_fit and n_parents                 pop_after_cross = Call crossover with pop_after_sel                 population_nextgen = Call mutation with pop_after_cross, mutation_rate, n_feat                 Append the best individual from generation i to best_chromo                 Append the best score from generation i to best_score             Endfor             Return best_chromo, best_score       Endfunction Calling the function     Gene,score = generations (X, y, size = 80, n_feat = X.shape[1], n_parents = 64, mutation_rate = 0.20, n_gen = 5, X_train = X_train, X_test = X_test, y_train = y_train, y_test = y_test)