def evaluate_accurancy(self, filename=''):
"""
Evaluate the Result of the emotion analyze
"""
from pycm import ConfusionMatrix
#actual_vector = self._data[CorporaProperties.EMOTION.value]
actual_vector = self._data.loc[:,
CorporaProperties.EMOTION.value].values
#print(actual_vector)
#predict_vector = self._data[CorporaProperties.CALCULATED_EMOTION.value]
predict_vector = self._data.loc[:, CorporaProperties.
CALCULATED_EMOTION.value].values
#print(predict_vector)
cm = ConfusionMatrix(actual_vector=actual_vector,
predict_vector=predict_vector)
#print(cm)
cm.save_csv(filename)
Xts,
yts,
# normalize='pred',
display_labels=labelNames,
xticks_rotation=40,
include_values=True,
values_format='d',
# values_format='.2f',
cmap='RdPu',
ax=ax2)
fname = savingPath + type(clf).__name__ + '_idx_{0}'.format(i)
saveFig(fig1, fname + 'predNormalized')
saveFig(fig2, fname + 'Direct')
cm = ConfusionMatrix(yts, y_pred)
cm.save_csv(fname)
with open(fname + '_overallStat.csv', 'w+', newline="") as csv_file:
writer = csv.writer(csv_file)
for key, value in cm.overall_stat.items():
writer.writerow([key, value])
# plt.figure()
# plt.hist(yts,bins=len(classes))
# thisClass=np.where(yts==i)[0]
#
#
#
# 1/0
#
X_test = np.load(input_testdata_path + "610" + ".npz")
y_test = np.load(output_testdata_path + "610" + ".npy")
X_test = X_test['arr_0']
print(np.shape(X_test))
data = model.predict_classes(X_test)
#print(np.shape(data))
#print(np.shape(y_test))
a, b = np.where(y_test == 1)
actual_array = np.empty(260)
predicted_array = np.empty(260)
actual_array = b
predicted_array = data
plot_model(
model,
show_shapes=True,
)
# Plot non-normalized confusion matrix
cm = ConfusionMatrix(actual_array, predicted_array)
print(cm)
cm.save_csv('confusion_matrix')