-
Notifications
You must be signed in to change notification settings - Fork 3
/
Copy pathexperiments.py
172 lines (155 loc) · 7.01 KB
/
experiments.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
165
166
167
168
169
170
171
172
import warnings
warnings.warn = lambda *args, **kwards: None # Supress warnings.
import sys
import pandas as pd
from sklearn.model_selection import train_test_split
from sklearn.preprocessing import LabelEncoder
from sklearn.utils import shuffle
from neo4j import ServiceUnavailable
from GraphOfDocs.neo4j_wrapper import Neo4jDatabase
from GraphOfDocs import utils
from GraphOfDocs import select
from prettytable import PrettyTable
from GraphOfDocs import config_experiments
from GraphOfDocs import evaluation
import timeit
results_table = PrettyTable(
['Method', 'Accuracy', 'Number of features',
'Train size', 'Test size', 'Details'])
evaluation_results = []
feature_selection_evaluation_results = []
start_time = timeit.default_timer()
print('')
print('%'*100)
print('!START OF THE EXPERIMENT!')
#print('DATASET DIR PATH: %s' % config_experiments.DATASET_PATH)
print(f'DATASET DIR PATH: {config_experiments.DATASET_PATH}')
print('MIN NUMBER OF DOCUMENTS PER SELECTED COMMUNITY: '
f'{config_experiments.MIN_NUMBER_OF_DOCUMENTS_PER_SELECTED_COMMUNITY}')
print('VARIANCE THRESHOLD: '
f'{config_experiments.VARIANCE_THRESHOLD}')
print('SELECT KBEST K: '
f'{config_experiments.SELECT_KBEST_K}')
print('TOP N SELECTED COMMUNITY TERMS: '
f'{config_experiments.TOP_N_SELECTED_COMMUNITY_TERMS}')
# Connect to database.
try:
database = Neo4jDatabase('bolt://localhost:7687', 'neo4j', '123')
# Neo4j server is unavailable.
# This client app cannot open a connection.
except ServiceUnavailable as error:
print('\t* Neo4j database is unavailable.')
print('\t* Please check the database connection before running this app.')
input('\t* Press any key to exit the app...')
sys.exit(1)
# Retrieve the communities of documents and their filenames.
doc_communities = select.get_communities_filenames(database)
# Keep only the communities with more than one documents.
filtered_doc_communities = \
[doc_community for doc_community in doc_communities
if doc_community[2] >=
config_experiments.MIN_NUMBER_OF_DOCUMENTS_PER_SELECTED_COMMUNITY]
# Fetch the selected documents.
selected_docs = sum([docs for _, docs, _ in filtered_doc_communities], [])
# Map community id to documents.
doc_communities_dict = {community_id: docs
for community_id, docs, number_of_docs
in filtered_doc_communities}
# Map document to community id.
doc_to_community_dict = {doc: community_id
for community_id, doc_community, _
in filtered_doc_communities for doc in doc_community}
print(f'Number of selected documents: {len(selected_docs)}')
# Read dataset, clean dataset and create a pandas dataframe of the dataset.
dataset = utils.read_dataset(config_experiments.DATASET_PATH)
# Create a label encoder (map classes to integer numbers).
le = LabelEncoder()
# The class of each document can be found by simply split (character '_') its filename. E.g. comp.sys.mac.hardware_51712.
le.fit([config_experiments.extract_file_class(file[0]) for file in dataset])
# Tuple: file identifier, file class, file class number, file text.
clean_dataset = [(file[0],
config_experiments.extract_file_class(file[0]),
le.transform([config_experiments.extract_file_class(file[0])])[0],
' '.join(utils.generate_words(
file[1],
extend_window = True,
insert_stopwords = False,
lemmatize = False, stem = False)))
for file in dataset]
df = pd.DataFrame(clean_dataset,
columns = ['identifier', 'class', 'class_number', 'text'])
df_all = df
# Keep only the selected documents (i.e. the document from the community with more than 1 documents).
df = df[df['identifier'].isin(selected_docs)]
df = shuffle(df, random_state = 42)
print('EXAMPLE OF THE PANDAS DATAFRAME')
print(df.head(2))
# Number of unique classes
print(f'Number of unique classes: {le.classes_.shape}')
X = df['text']
y = df['class_number']
positions = [i for i in range(len(X))]
positions_train, positions_test = train_test_split(
positions, test_size =0.33, random_state = 42)
res = evaluation.BOWEvaluator()\
.evaluate(X, y, results_table = results_table,
classifiers = config_experiments.classifiers)
evaluation_results.extend(res)
res = evaluation.MetaFeatureSelectionEvaluator()\
.evaluate(X, y, results_table = results_table,
classifiers = config_experiments.classifiers)
evaluation_results.extend(res)
for variance_threshold in config_experiments.VARIANCE_THRESHOLD:
res = evaluation\
.LowVarianceFeatureSelectionEvaluator(
variance_threshold=variance_threshold)\
.evaluate(X, y, results_table = results_table,
classifiers = config_experiments.classifiers)
evaluation_results.extend(res)
feature_selection_evaluation_results.extend(res)
for kbest_k in config_experiments.SELECT_KBEST_K:
res = evaluation\
.SelectKBestFeatureSelectionEvaluator(
kbest=kbest_k)\
.evaluate(X, y, results_table = results_table,
classifiers=config_experiments.classifiers)
evaluation_results.extend(res)
feature_selection_evaluation_results.extend(res)
evaluation.GraphOfDocsClassifier(
doc_to_community_dict, doc_communities_dict)\
.calculate_accuracy(df['identifier'], results_table = results_table)
for top_n in config_experiments.TOP_N_SELECTED_COMMUNITY_TERMS:
res = evaluation\
.TopNOfEachCommunityEvaluator(
top_n, doc_to_community_dict,
doc_communities_dict)\
.evaluate(
X, y, df = df,
positions_train = positions_train,
database = database,
results_table = results_table,
classifiers = config_experiments.classifiers)
evaluation_results.extend(res)
feature_selection_evaluation_results.extend(res)
df_evaluation_results = pd.DataFrame(evaluation_results)
df_feature_selection_evaluation_results = \
pd.DataFrame(feature_selection_evaluation_results)
print('EXAMPLE OF THE EVALUATION RESULTS PANDAS DATAFRAME')
print(df_evaluation_results.head(2))
results_table.sortby = 'Accuracy'
results_table.reversesort = True
print(results_table)
output_dir = config_experiments.EXPERIMENTAL_RESULTS_OUΤPUT_DIR
plots_prefix = config_experiments.PLOTS_PREFIX
df_evaluation_results.to_csv(f'{output_dir}/{plots_prefix}_evaluation_results.csv')
evaluation.generate_plots(
df_feature_selection_evaluation_results,
output_dir = output_dir,
plots_prefix = f'{plots_prefix}_feature_selection',
show_only = False)
database.close()
stop_time = timeit.default_timer()
print(f'Execution time: {stop_time - start_time}')
print('!END OF THE EXPERIMENT!')
print('%'*100)
print('')