def run_classifier(files: Tuple[str, str,
str], method: str, method_class: Base,
model_file: str, lower_case: bool) -> None:
"Inherit classes from classifiers.py and apply the predict/accuracy methods"
train, dev, test = files # Unpack train, dev and test filenames
result = method_class.predict(train, test, lower_case)
method_class.accuracy(result)
# Plot confusion matrix
make_dirs("Plots")
fig, ax = plot_confusion_matrix(result['truth'],
result['pred'],
normalize=True)
ax.set_title("Normalized Confusion Matrix: {}".format(method.title()))
fig.tight_layout()
fig.savefig("Plots/{}.png".format(method))
if not os.path.exists(directory):
os.makedirs(directory)
os.makedirs(directory + "/normalized")
os.makedirs(directory + "/unnormalized")
model.save(directory + "/" + file_name +
".model")
i += 1
# print(model.evaluate(x_test, y_test, batch_size=batch_size))
(cnf_mat, acc) = gue.manual_verification_100(
model, (x_test, y_test), batch_size=b_size)
plt.figure(figsize=(10, 10), dpi=100)
plotter.plot_confusion_matrix(
cnf_mat,
classes=range(16),
normalize=True,
title='Normalized confusion matrix')
plt.savefig(directory + "/normalized/" +
file_name + "_normalized_" +
str(acc) + ".png")
plt.figure(figsize=(10, 10), dpi=100)
plotter.plot_confusion_matrix(
cnf_mat.astype(int),
classes=range(16),
normalize=False,
title='Non-Normalized confusion matrix')
plt.savefig(directory + "/unnormalized/" +
file_name + "_" + str(acc) +
def run(args):
"""Executes the main process of the script
Parameters
----------
args : ArgumentParser
The arguments of the command typed by the user
"""
global CONFIG
CONFIG = exh.load_json("config/{0}.json".format(args.CONFIG))
print("Loading publications")
# Load DIDA publications
dida_data = exh.load_json(FILENAME_TEMPLATE.format(CONFIG['DIDA_DOCS']))
# Load Not-DIDA publications
notdida_data = exh.load_json(
FILENAME_TEMPLATE.format(CONFIG['NOTDIDA_DOCS']))
# docs = [deepcopy(notdida_data), deepcopy(dida_data)]
docs = [deepcopy(dida_data), deepcopy(notdida_data)]
display.display_ok("Loading publications done")
data_directory = DIRECTORY + '/' + CONFIG['ALL_CLUSTERS_DIRECTORY']
Ndw = exh.load_json(data_directory + "/ndw.json")
W = exh.load_json(data_directory + "/W.json")
# Real labels of each publication
# y_true = np.append(np.zeros(len(notdida_data)), np.ones(len(dida_data)))
y_true = np.append(np.ones(len(dida_data)), np.zeros(len(notdida_data)))
strict_result, doublon_result = classification(docs, Ndw, W,
data_directory, y_true)
plt.plot_confusion_matrix(strict_result,
len(dida_data),
len(notdida_data),
"strict_",
"n_clusters",
"Number of clusters",
DIRECTORY,
step=1000)
exh.save_to_log(strict_result, "strict", "n_clusters",
LOG_FILENAME.format("strict"))
plt.plot_confusion_matrix(doublon_result,
len(dida_data),
len(notdida_data),
"doublon_",
"n_clusters",
"Number of clusters",
DIRECTORY,
step=1000)
exh.save_to_log(doublon_result, "doublon", "n_clusters",
LOG_FILENAME.format("doublon"))
scores = [strict_result['score'], doublon_result['score']]
classifiers_names = ["Strict converter", "Doublon converter"]
plt.plot_lines(strict_result['n_clusters'],
scores,
classifiers_names,
FSCORE_FILENAME,
"Number of clusters",
"F1-score",
step=1000)
def run(args):
"""Executes the main process of the script
Parameters
----------
args : ArgumentParser
The arguments of the command typed by the user
"""
global CONFIG
CONFIG = exh.load_json("config/{0}.json".format(args.CONFIG))
exh.create_directory(DIRECTORY)
print("Loading publications")
# Load DIDA publications
dida_data = exh.load_json(FILENAME_TEMPLATE.format(CONFIG['DIDA_DOCS']))
# Load Not-DIDA publications
notdida_data = exh.load_json(
FILENAME_TEMPLATE.format(CONFIG['NOTDIDA_DOCS']))
display.display_ok("Loading publications done")
n = CONFIG['NGRAMS']
csv_files = csv_filenames(n)
# Real labels of each publication
y_true = np.append(np.ones(len(dida_data)), np.zeros(len(notdida_data)))
data = deepcopy(dida_data)
data.extend(deepcopy(notdida_data))
scores = []
classifiers_names = []
print("Strict Classifier training")
results = train(StrictClassifier, deepcopy(data), csv_files, y_true)
plt.plot_confusion_matrix(results, len(dida_data), len(notdida_data),
'strict_', "threshold", "Threshold", DIRECTORY)
scores.append(results['score'])
exh.save_to_log(results, "strict", "threshold",
LOG_FILENAME.format("strict"))
classifiers_names.append("Strict Classifier")
display.display_ok("Strict Classifier training done")
print("Split Weighted Classifier training")
results = train(SplitWeightedClassifier, deepcopy(data), csv_files, y_true)
plt.plot_confusion_matrix(results, len(dida_data), len(notdida_data),
'splitweighted_', "threshold", "Threshold",
DIRECTORY)
scores.append(results['score'])
exh.save_to_log(results, "splitweighted", "threshold",
LOG_FILENAME.format("splitweighted"))
classifiers_names.append("Split Weighted Classifier")
display.display_ok("Split Weighted Classifier training done")
print("Weighted Classifier training")
results = train(WeightedClassifier, deepcopy(data), csv_files, y_true)
plt.plot_confusion_matrix(results, len(dida_data), len(notdida_data),
'weighted_', "threshold", "Threshold", DIRECTORY)
scores.append(results['score'])
exh.save_to_log(results, "weighted", "threshold",
LOG_FILENAME.format("weighted"))
classifiers_names.append("Weighted Classifier")
display.display_ok("Weighted Classifier training done")
plt.plot_lines(results['threshold'], scores, classifiers_names,
FSCORE_FILENAME, "Threshold", "F1-score")
display.display_info("Results saved in {0}".format(DIRECTORY))