def csvConversion(old_filepath, new_filepath):
# Open and read txt file under variable "file"
file = open(old_filepath, "r")
csv = csvFileEvaluation()
#Create csv with identified .txt file rows and columns
csv.createCSV(file, old_filepath, new_filepath)
#Print rows and columnes of "file"
print("\nNew .csv file dimensions: ")
print("Rows: ", csv.rows)
print("Columns: ", csv.columns)
# Close .txt file
file.close()
y_train, y_test = Y[train_index], Y[test_index]
# Naive Bayes
elapsedTraining, elapsedTesting, y_pred = nbClassifier(X_train, y_train, X_test)
# Calculate the different accuracy scores
cm_metrics = Calculate_Metrics(y_test, y_pred)
metrics_dict = cm_metrics.calculateScore()
# Calculate the run time
elapsed = (timeit.default_timer() - start_time) * 1000
metrics_dict.update({'Run Time(MSec)': elapsed})
# Create metrics for each fold
pdf_indi = printPDF(metrics_dict)
pdf_indi.create_pdf_indi(test_spec_dict, test_description, test_path_indi,
str(fold) + '_' + testTimeConvertet + '_' + hostname)
metrics_array.append(metrics_dict)
# Calculate the overall score for all the folds
metrics_mean_dict = cm_metrics.calculate_mean_score(metrics_array, folds)
# # Create PDF
pdf = printPDF(metrics_mean_dict)
pdf.create_pdf(test_spec_dict, test_description, test_path, '_' + testTimeConvertet + '_' + hostname)
# Create CSV
csv = printCSV(metrics_mean_dict, test_spec_dict)
csv.createCSV(path, csvfile_name)
def runTest(self,
path,
treeSize,
trainingFile,
test_description,
type,
traning_size,
create_pdf=False):
## Overall configurations
hostname = socket.gethostname()
#testStart = datetime.datetime.now()
#testStartConvertet = time.strftime("(%Y-%m-%d)-(%H-%M-%S)")
# path = testStartConvertet + '_' + hostname + '/'
try:
os.makedirs(path)
except OSError:
if not os.path.isdir(path):
raise
## Induvidual test configurations
testTime = datetime.datetime.now()
testTimeConvertet = time.strftime("(%Y-%m-%d)-(%H-%M-%S)")
csvfile_name = str(treeSize) + '_' + testTimeConvertet + '_' + str(
traning_size) + '.csv'
test_path = path + str(treeSize) + '/'
test_path_indi = path + str(treeSize) + '/' + 'Folds/'
try:
os.makedirs(test_path)
except OSError:
if not os.path.isdir(test_path):
raise
try:
os.makedirs(test_path_indi)
except OSError:
if not os.path.isdir(test_path_indi):
raise
print "DT Test: " + str(treeSize) + " Time: " + testTimeConvertet
# Training Set Path setup
traning_set = trainingFile
# Load training set
test_set = loadTestSet(traning_set)
X, Y = test_set.loadTestSet()
# Set number of folds
folds = 6
kf = KFold(n_splits=folds)
kf.get_n_splits(X)
metrics_array = []
fold = 0
test_spec_dict = {
'Classifier': 'DT',
'Depth': treeSize,
'Date': testTimeConvertet,
'Host': hostname,
'Training Set': traning_set,
'Sample Type': type
}
elapsedTraining1 = 0
elapsedTesting2 = 0
for train_index, test_index in kf.split(X):
fold += 1
start_time = timeit.default_timer()
X_train, X_test = X[train_index], X[test_index]
y_train, y_test = Y[train_index], Y[test_index]
# Classifies fields with different k values from 1 to 50.
elapsedTraining, elapsedTesting, y_pred = dtClassifier(
treeSize, X_train, y_train, X_test)
elapsedTraining1 += elapsedTraining
elapsedTesting2 += elapsedTesting
# Calculate the different accuracy scores
cm_metrics = Calculate_Metrics(y_test, y_pred)
metrics_dict = cm_metrics.calculateScore()
# Calculate the run time
elapsed = (timeit.default_timer() - start_time) * 1000
metrics_dict.update({'Run Time(MSec)': elapsed})
if create_pdf:
# Create metrics for the fold
pdf_indi = printPDF(metrics_dict)
pdf_indi.create_pdf_indi(
test_spec_dict, test_description, test_path_indi,
str(fold) + '_' + testTimeConvertet + '_' + hostname)
metrics_array.append(metrics_dict)
# Calculate the overall score for all the folds
metrics_mean_dict = cm_metrics.calculate_mean_score(
metrics_array, folds)
print 'Time in ms for training: ' + str(elapsedTraining1 / 6)
print 'Time in ms for testing: ' + str(elapsedTesting2 / 6)
if create_pdf:
# Create PDF
pdf = printPDF(metrics_mean_dict)
pdf.create_pdf(
test_spec_dict, test_description, test_path,
str(treeSize) + '_' + testTimeConvertet + '_' + hostname)
# Create CSV
csv = printCSV(metrics_mean_dict, test_spec_dict)
csv.createCSV(path, csvfile_name)
return csvfile_name