Feature-Selection-for-Text/Algorithm_Fcs_trial.py at master · TheCaesar13/Feature-Selection-for-Text · GitHub

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
import numpy as np
from statistics import median
import math
import time
import random
import statistics

def sigmoid1(gamma):
    # print(gamma)
    if gamma < 0:
        return 1 - 1 / (1 + math.exp(gamma))
    else:
        return 1 / (1 + math.exp(-gamma))

def fitness(total_chi, pop_size, dimension, search_agent_matrix,sum_fitness_SA, fitness_SA, chi_val):

    number_of_classes = len(chi_val)
    all_fitness = np.zeros([pop_size, number_of_classes])
    for n in range(0, pop_size):
        for i in range(0, dimension):
            if search_agent_matrix[n][i] == 1:
                fitness_SA.append(chi_val[0:number_of_classes,i])
        # put all search agents fitnesses in one array and sort it
        #all_fitness.append(statistics.harmonic_mean(fitness_SA))
        print("Shape of fitness matrix for an agent", np.shape(fitness_SA))
        #    SUM THE CHI VALUES FOR EVERY CLASS
        for i in range(0, number_of_classes):
            sum_fitness_SA.append(sum(fitness_SA[:][i]))
        all_fitness[n] = sum_fitness_SA
        print("Sumed fitness for each class", sum_fitness_SA)
        fitness_SA.clear()
        sum_fitness_SA.clear()
    return all_fitness


def algorithm_fcs(total_chi, iter_nr, pop_size, chi_val, dimension):
    # initialize position of search agents/population
    """random matrix of (3-search agents,1144-words in the corpus/features)"""
    search_agent_matrix = np.zeros([pop_size, dimension])

    for i in range(0, pop_size):
        for j in range(0, dimension):
           search_agent_matrix[i][j] = int(random.uniform(0,2))
    #print("First agent \n", np.shape(search_agent_matrix[0:1, :]), "\n")
    iteration = 1
    best_ever = 0
    best_position = []
    sum_fitness_SA = []
    fitness_SA = []
    #all_fitness = []
    while iteration < iter_nr + 1:

        print("Current Iteration: ", iteration)
        # calculate the fitness of each individual
        all_fitness = fitness(total_chi, pop_size, dimension, search_agent_matrix,sum_fitness_SA, fitness_SA, chi_val)
        # put all search agents fitnesses in one array and sort it
        print("Search agent fitness unordered", all_fitness)
        all_fitness_sorted = all_fitness
        for n in range(0, pop_size-1):
            for m in range(0, pop_size-n-1):
                if(sum(np.subtract(all_fitness_sorted[m], all_fitness_sorted[m+1])) > 0):
                    #all_fitness_sorted[m], all_fitness_sorted[m+1] = all_fitness_sorted[m + 1], all_fitness_sorted[m]
                    all_fitness_sorted[m] = all_fitness_sorted[m] * all_fitness_sorted[m + 1]
                    all_fitness_sorted[m + 1] = all_fitness_sorted[m] / all_fitness_sorted[m + 1]
                    all_fitness_sorted[m] = all_fitness_sorted[m] / all_fitness_sorted[m + 1]
        print("Ascending order of search agents by fitness:", all_fitness_sorted)
        #print("Ascending order of search agents by fitness:", all_fitness_sorted)
        # OBTAIN AND STORE THE BEST ALL TIME INDIVIDUAL
        best_fitness = all_fitness[pop_size-1]
        print("Best fitness", best_fitness)
        print("All fitness length",len(all_fitness))
        if sum(np.subtract(best_ever, best_fitness)) < 0:
            best_ever = best_fitness
        print("Best ever", best_ever)
        for i in range(0, len(all_fitness)):
            if sum(np.subtract(all_fitness[i], best_fitness)) == 0:
                if iteration == 1:
                    best_position = search_agent_matrix[i][:]
                    break
                elif sum(np.subtract(best_fitness, best_ever)) > 0:
                    best_position = search_agent_matrix[i][:]
                    break
        print("Best position", best_position)
        worst_fitness = all_fitness[0]
        print("Worst fitness", worst_fitness)

        #print("all fitness:", all_fitness)
        # calculate the parameter "a"
        a = np.arctanh(-(iteration / iter_nr) + 1)
        b = 1 - iteration / iter_nr
        c = 0 + iteration / (4*iter_nr)
        # calculate the fitness weight of each slime mold
        weight = np.zeros([pop_size])
        for i in range(0, pop_size):
            """pow(2,-52) to avoid denominator 0"""
            fitness_formula_replacer = sum(np.subtract(best_fitness, all_fitness_sorted[i])) / (
                        sum(np.subtract(best_fitness, worst_fitness)) + pow(2, -52)) + 1
            print("Fitness formula shortcut", fitness_formula_replacer)

            if i >= (1/2*pop_size):
                weight[i] = (1 + np.random.uniform() * np.log10(fitness_formula_replacer))
                print("Strong weight", weight[i])
            else:
                weight[i] = (1 - np.random.uniform() * np.log10(fitness_formula_replacer))
                print("Week weight", weight[i])
            #print("fitness formula", fitness_formula_replacer)
            # UPDATE THE POSITIONS OF THE SEARCH AGENTS
            # calculate p parameter
        for i in range(0,pop_size):
            p = np.tanh(abs(median(np.subtract(all_fitness_sorted[i], best_ever))))
            # calculate vb and vc
            vb = random.uniform(-a, a)
            vc = random.uniform(-b, b)
            #print(len(search_agent_matrix[0][:]))
            A = random.randint(0, pop_size-1)
            B = random.randint(0, pop_size-1)
            r = random.random()
            if r < p:
                for j in range(0, dimension):
                    search_agent_matrix[i][j] = best_position[j] + vb * (int(weight[i]) * search_agent_matrix[A][j]
                                                                  - search_agent_matrix[B][j])
            else:
                for j in range(0, dimension):
                    search_agent_matrix[i][j] = vc * search_agent_matrix[i][j]


            for j in range(0, dimension):
                random.seed(time.time() * 200 + 999)
                r1 = random.random()
                if sigmoid1(search_agent_matrix[i][j]) > r1:
                    search_agent_matrix[i][j] = 1
                else:
                    search_agent_matrix[i][j] = 0
            print("Random value in sigmoid function:", r1)
        #print("best", best_position)
        for i in range(0, len(all_fitness)):
            if sum(np.subtract(all_fitness[i], best_fitness)) == 0:
                if iteration == 1:
                    best_position = search_agent_matrix[i][:]
                    break
                elif sum(np.subtract(best_fitness, best_ever)) > 0:
                    best_position = search_agent_matrix[i][:]
                    break
        print("Best position", best_position)

        iteration += 1
        """
        print("First fitness \n", fitness_SA1, "\n")
        print("Second fitness \n", fitness_SA2, "\n")
        print("Third fitness \n", fitness_SA3, "\n")
        """
        #print("These are the weights: ", weight)
    #print("last position", search_agent_matrix)
        counter = 0
        anti_counter = 0
        for i in range(0, dimension):
            if best_position[i] == 1:
                counter += 1
            elif best_position[i] == 0:
                anti_counter += 1
        print("Solution 1: ", counter, "and 0: ", anti_counter)


    return best_position