-
Notifications
You must be signed in to change notification settings - Fork 0
/
GraphsCreation.py
164 lines (152 loc) · 7.6 KB
/
GraphsCreation.py
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 matplotlib.pyplot as plt
import csv
import re
import numpy as np
class GraphCreation:
def __init__(self, max_level, data, classifiers, vectorization_types, selected_parameters, param_selection_folder, graphs_folder):
print("GraphCreation: starts")
self.param_selection_folder = param_selection_folder
self.graphs_folder = graphs_folder
self.classifier_names = [classifier.name for classifier in classifiers]
self.category_names = []
self.category_ids = []
self.selected_parameters = selected_parameters
for level in range(max_level+1):
for category_id in data.categories_ids_by_levels[level]:
self.category_ids.append(category_id)
self.category_names.append(data.categories[category_id].title.replace('category:', ''))
self.vectorization_types_names = [str(vec_type).replace('VectorizationType.', '') for vec_type in vectorization_types]
self.vectorization_types = vectorization_types
self.parameters = {}
for clf_name in self.classifier_names:
self.parameters[clf_name] = {}
for category_name in self.category_names:
self.parameters[clf_name][category_name] = {}
self.load_files()
self.create_basic_graphs()
self.create_accuracy_histograms()
print("GraphCreation: finished")
def get_path(self, clf_name, vectorization_type, category_name):
return self.param_selection_folder + clf_name + "_VectorizationType." + vectorization_type + "_category." + category_name \
+ ".csv"
def load_files(self):
for clf_name in self.classifier_names:
for category_name in self.category_names:
for vectorization_type in self.vectorization_types_names:
file_path = self.get_path(clf_name, vectorization_type, category_name)
with open(file_path, mode='r') as file:
reader = csv.reader(file)
headers = next(reader,None)[1:]
columns = {}
for h in headers:
columns[h] = []
for row in reader:
row = row[1:]
for h,v in zip(headers,row):
columns[h].append(v)
self.parameters[clf_name][category_name][vectorization_type] = columns
def get_best_parameters_for(self, clf_name, category_name):
max_score = -1
best_params = {}
current_params = self.parameters[clf_name][category_name]
for vec_type in self.vectorization_types_names:
for s in current_params[vec_type]['score']:
if float(s) > max_score:
max_score = float(s)
s_index = current_params[vec_type]['score'].index(s)
for key in current_params[vec_type].keys():
if key != 'score':
best_params[key] = current_params[vec_type][key][s_index]
return best_params
def get_column_for_free_param(self, fixed_params, free_param, clf_name, category_name, vec_type):
columns = {'score': [], free_param: []}
current_params = self.parameters[clf_name][category_name][vec_type]
for i in range(len(current_params['score'])):
should_take_row_i = True
for key in fixed_params:
if current_params[key][i] != fixed_params[key]:
should_take_row_i = False
break
if should_take_row_i:
columns['score'].append(float(current_params['score'][i]))
if re.search('[a-zA-Z]', current_params[free_param][i]):
columns[free_param].append(current_params[free_param][i])
else:
columns[free_param].append(float(current_params[free_param][i]))
return columns
def create_graphs_for(self, clf_name, category_name):
best_params = self.get_best_parameters_for(clf_name,category_name)
for free_param in best_params.keys():
fixed_params = dict(best_params)
fixed_params.pop(free_param, None)
plt.xlabel(beautify_string(free_param))
plt.ylabel(beautify_string('score'))
for vec_type in self.vectorization_types_names:
graph_series = self.get_column_for_free_param(fixed_params, free_param, clf_name, category_name, vec_type)
plt.plot(graph_series[free_param], graph_series['score'])
plt.legend([beautify_string(vec_type) for vec_type in self.vectorization_types_names])
plt.title(beautify_string(category_name), pad=30, fontsize=18, x=0.44)
fixed_params_string = ": "
for best_param_name in best_params.keys():
if best_param_name != free_param:
fixed_params_string += beautify_string(best_param_name) + " = " + best_params[best_param_name] + " "
if fixed_params_string == ": ":
fixed_params_string = ""
plt.suptitle(beautify_string(clf_name) + fixed_params_string,fontsize=11,y=0.89)
plt.grid(linewidth=0.2)
plt.tight_layout()
plt.savefig(self.graphs_folder + clf_name + "_" + free_param + "_" + category_name + '.png', dpi=300)
plt.clf()
def create_basic_graphs(self):
for clf_name in self.classifier_names:
for category_name in self.category_names:
self.create_graphs_for(clf_name, category_name)
def create_accuracy_histograms(self):
for category_id in self.category_ids:
self.create_accuracy_histograms_for(category_id)
def create_accuracy_histograms_for(self, category_id):
best_parameters = self.selected_parameters
x = np.arange(len(self.classifier_names))
width = 0.25
x_coordinates = [x-width, x, x+width]
fig, ax = plt.subplots()
for vec_type in self.vectorization_types:
y_values = []
for clf_name in self.classifier_names:
y_values.append(best_parameters[str((clf_name, vec_type, category_id))][1])
ax.bar(x_coordinates.pop(), y_values, width, label=beautify_string(str(vec_type)))
plt.yticks(np.arange(0.4, 1.05, 0.05))
plt.ylim(0.4, 1)
ax.set_xticks(x)
classifier_names = [beautify_string(classifier_name, is_for_histogram_axis=True) for classifier_name in
self.classifier_names]
ax.set_xticklabels(classifier_names)
ax.set_ylabel("Accuracy")
category_name = beautify_string(self.category_names[int(category_id)])
plt.title(category_name, fontsize=18, pad=10, x=0.47)
plt.grid(linewidth=0.2)
ax.legend()
plt.savefig(self.graphs_folder + category_name + '_accuracy_histogram.png', dpi=300)
plt.clf()
def beautify_string(string, is_for_histogram_axis=False):
if is_for_histogram_axis:
if string == 'KNN':
return 'k-NN'
if string == 'Id3':
return 'ID3'
string = string.replace('_', '\n')
string = string.replace(' ', '\n')
return string
string = string.replace('nn__', '')
string = string.replace('_', ' ')
string = string.replace('VectorizationType.', '')
string = string.title()
if string == "Svm":
return "SVM"
if string == "Knn":
return "k-Nearest Neighbors"
if string == "N Estimators":
return "N-Estimators"
if string == "N Neighbors":
return "N-Neighbors"
return string