output_files = []
output_files.append("file,mse,parameters")
helpers.dump_results(output_files, results_folder,model)
for f in csv_input_files:
try :
# fetching input file data
dataframe_expect = pandas.read_csv(f)
value = np.array(dataframe_expect['value'])
timestamp = np.array(dataframe_expect['timestamp'])
try :
label = np.array(dataframe_expect['label'])
except KeyError:
print "Warnning :["+f+"] No `label` colomn found data set without labels !. Assumes there are no anomolies"
label = np.zeros(len(value))
dataframe_predction = pandas.read_csv(helpers.get_result_file_name(f,results_folder,model))
prediction = np.array(dataframe_predction['prediction'])
params = "training_size="+str(tesing_start)
file_not_found = False
except IOError:
file_not_found = True
print("Warnning :["+helpers.get_result_file_name(f,results_folder,model)+"] File not found !")
if file_not_found:
out_file = helpers.get_result_file_name(f, results_folder,model)
output_files.append( helpers.get_result_dump_name(out_file) +",n/a,n/a" )
helpers.dump_results(output_files, results_folder,model)
else :
training = np.ones(tesing_start)
testing = np.zeros(len(value)-tesing_start)
data = {
'value': value,
'prediction_training': training_colomn,
'label': label
}
colomn_arrangement = ['value', 'prediction_training', 'label']
for i in range(no_of_prediction_points):
prediction[i] = np.append(training, prediction[i])
data['prediction_' + str(i)] = prediction[i]
colomn_arrangement.append('prediction_' + str(i))
dataframe_out = pandas.DataFrame(data, index=timestamp)
dataframe_out.index.name = "timestamp"
dataframe_out = dataframe_out[colomn_arrangement]
out_file = helpers.get_result_file_name(f, csv_output_directory, m)
dataframe_out.to_csv(out_file)
testing_value = value[testing_start:]
mse = helpers.MSE_multipoint(testing_value, testing_prediction,
no_of_prediction_points)
output_files.append(
helpers.get_result_dump_name(out_file) + "," + str(mse) + "," +
params)
helpers.dump_results(output_files, csv_output_directory, m)
print("##### [" + m + "]" + str(count) +
" CSV input File processed #####")
count += 1
print("##### " + m + " done ! #####")
output_files.append("file,mse,parameters")
helpers.dump_results(output_files, results_folder, model)
for f in csv_input_files:
try:
# fetching input file data
dataframe_expect = pandas.read_csv(f)
value = np.array(dataframe_expect['value'])
timestamp = np.array(dataframe_expect['timestamp'])
try:
label = np.array(dataframe_expect['label'])
except KeyError:
print "Warnning :[" + f + "] No `label` colomn found data set without labels !. Assumes there are no anomolies"
label = np.zeros(len(value))
dataframe_predction = pandas.read_csv(
helpers.get_result_file_name(f, results_folder, model))
prediction = np.array(dataframe_predction['prediction'])
params = "training_size=" + str(tesing_start)
file_not_found = False
except IOError:
file_not_found = True
print("Warnning :[" +
helpers.get_result_file_name(f, results_folder, model) +
"] File not found !")
if file_not_found:
out_file = helpers.get_result_file_name(f, results_folder, model)
output_files.append(
helpers.get_result_dump_name(out_file) + ",n/a,n/a")
helpers.dump_results(output_files, results_folder, model)
else: