MAGICGrants/Monero-Dataset-Pipeline
1
0
Fork 0
mirror of https://github.com/MAGICGrants/Monero-Dataset-Pipeline.git synced 2026-01-09 13:37:57 -05:00
Code Issues Packages Projects Releases Wiki Activity

- fixed NN plots

Browse Source
This commit is contained in:
ACK-J
2022-08-14 11:54:31 -04:00
parent eca377f1f9
commit 4e6db7f16c
1 changed files with 9 additions and 10 deletions
Download Patch File Download Diff File Expand all files Collapse all files

19
DataScience/NeuralNetwork.py
View File

@@ -89,20 +89,20 @@ def MLP(X_train, X_test, y_train, y_test, X_Validation, y_Validation, stagenet=T
scaler = StandardScaler().fit(X_Validation)
X_Validation = scaler.transform(X_Validation)
# fix off by one error
for i in range(len(y_test_copy)):
y_test_copy[i] = y_test_copy[i] - 1
for i in range(len(y_val_copy)):
y_val_copy[i] = y_val_copy[i] - 1
for i in range(10):
from keras_visualizer import visualizer
model = Sequential()
model.add(Dense(11, input_shape=(X_train.shape[1],), activation='relu'))
#model.add(Dense(32, activation='relu'))
model.add(Dense(64, activation='relu'))
#model.add(Dense(128, activation='relu'))
model.add(Dropout(.1))
#model.add(Dense(256, activation='relu'))
#model.add(Dense(128, activation='relu'))
model.add(Dense(64, activation='relu'))
#model.add(Dropout(.3))
model.add(Dense(32, activation='relu'))
#model.add(Dropout(.2))
model.add(Dense(11))
model.add(BatchNormalization())
model.add(Activation('softmax'))
@@ -133,9 +133,7 @@ def MLP(X_train, X_test, y_train, y_test, X_Validation, y_Validation, stagenet=T
score = model.evaluate(X_test, y_test, verbose=1)
print(score[1])
y_pred = list(model.predict(X_Validation).argmax(axis=1))
for i in range(len(y_pred)):
y_pred[i] = y_pred[i] + 1
# y_pred = list(model.predict(X_Validation).argmax(axis=1))
#confusion_mtx(y_Validation, y_pred)
# Metrics
@@ -145,9 +143,9 @@ def MLP(X_train, X_test, y_train, y_test, X_Validation, y_Validation, stagenet=T
weighted_f1 = f1_score(y_test_copy, y_pred, average='macro')
print('Weighted F1-score: {:.2f}'.format(weighted_f1))
out_of_sample_f1.append(weighted_f1)
if stagenet:
cm = confusion_matrix(y_test_copy, y_pred)
cm = cm.astype('float') / cm.sum(axis=1)[:, np.newaxis]
# Heat map
plt.figure(figsize=(10, 7))
sn.heatmap(cm, annot=True)
@@ -156,6 +154,7 @@ def MLP(X_train, X_test, y_train, y_test, X_Validation, y_Validation, stagenet=T
plt.savefig("./models/NN/stagenet/CM_epochs_" + str(EPOCHS) + "_batch_size_" + str(BATCH_SIZE) + "_i_" + str(i) + "_accuracy_" + str(weighted_f1) + ".png")
else:
cm = confusion_matrix(y_test_copy, y_pred)
cm = cm.astype('float') / cm.sum(axis=1)[:, np.newaxis]
# Heat map
plt.figure(figsize=(10, 7))
sn.heatmap(cm, annot=True)