print("F1 Score: {}".format(f1_score(y_pred_m, y_test_m, average = 'weighted')))
print("Precision: {}".format(precision_score(y_pred_m, y_test_m, average='weighted')))
print("Recall: {}".format(recall_score(y_pred_m, y_test_m, average='weighted')))
#Performance Report, Motions
for particle_type in class_names_p:
pred_score_p = best_p.score(X_test_p[y_test_p.id==particle_type], y_test_p[y_test_p.id==particle_type])
print("{} accuracy = {p:8.4f}".format(particle_type, p=pred_score_p))
print("Cohen Kappa: {}".format(cohen_kappa_score(y_pred_p, y_test_p)))
print("Accuracy: {}".format(accuracy_score(y_pred_p, y_test_p)))
print("F1 Score: {}".format(f1_score(y_pred_p, y_test_p, average = 'weighted')))
print("Precision: {}".format(precision_score(y_pred_p, y_test_p, average='weighted')))
print("Recall: {}".format(recall_score(y_pred_p, y_test_p, average='weighted')))
learning_curves = False
if learning_curves:
plot_learning_curves(model_m, X_train1_m, y_train1_m, low_limit=0.5, title = "Decision Tree - Motions Set - Post-Tuning Learning Curves")
plot_learning_curves(model_p, X_train1_p, y_train1_p, low_limit=0.5, title = "Decision Tree - Particles Set - Post-Tuning Learning Curves")
print("time elapsed: {}".format(time.time()-t0))
# References
# https://medium.com/@rnbrown/creating-and-visualizing-decision-trees-with-python-f8e8fa394176
# https://towardsdatascience.com/decision-trees-pruning-4241cc266fef
# https://stats.stackexchange.com/questions/28029/training-a-decision-tree-against-unbalanced-data
# https://www.geeksforgeeks.org/decision-tree-implementation-python/
print("Cohen Kappa: {}".format(cohen_kappa_score(y_pred, y_test)))
print("Accuracy: {}".format(accuracy_score(y_pred, y_test)))
print("Balanced Accuracy: {}".format(
balanced_accuracy_score(y_pred, y_test)))
print("F1 Score: {}".format(f1_score(y_pred, y_test, average='weighted')))
print("Precision: {}".format(
precision_score(y_pred, y_test, average='weighted')))
print("Recall: {}".format(recall_score(y_pred, y_test,
average='weighted')))
learning_curves = False
if learning_curves:
estimator = KerasClassifier(build_fn=classification_model,
epochs=epo,
batch_size=bat,
verbose=0)
#scorer = make_scorer(cohen_kappa_score)
plot_learning_curves(
estimator,
X_train1,
y_train1,
title="Neural Network - Particles Set - Post-Tuning Learning Curves",
low_limit=0.6)
print("time elapsed: {}".format(time.time() - t0))
#References:
# borrowed heavily from
# https://www.tensorflow.org/tutorials/keras/basic_classification
# https://machinelearningmastery.com/multi-class-classification-tutorial-keras-deep-learning-library/
# https://machinelearningmastery.com/display-deep-learning-model-training-history-in-keras/
pred_score_p = best_p.score(X_test_p[y_test_p.id == particle_type],
y_test_p[y_test_p.id == particle_type])
print("{} accuracy = {p:8.4f}".format(particle_type, p=pred_score_p))
print("Cohen Kappa: {}".format(cohen_kappa_score(y_pred_p, y_test_p)))
print("Accuracy: {}".format(accuracy_score(y_pred_p, y_test_p)))
print("F1 Score: {}".format(
f1_score(y_pred_p, y_test_p, average='weighted')))
print("Precision: {}".format(
precision_score(y_pred_p, y_test_p, average='weighted')))
print("Recall: {}".format(
recall_score(y_pred_p, y_test_p, average='weighted')))
learning_curves = False
if learning_curves:
plot_learning_curves(
model_m,
X_train1_m,
y_train1_m,
low_limit=0.6,
title="SVM - Motions Set - Post-Tuning Learning Curves")
plot_learning_curves(
model_p,
X_train1_p,
y_train1_p,
low_limit=0.6,
title="SVM - Particles Set - Post-Tuning Learning Curves")
print("time elapsed: {}".format(time.time() - t0))
#References
#https://scikit-learn.org/stable/modules/svm.html