def main():
data_loc = 'no_name_counts_actresses.txt'
samples_per_career = 5
min_length = 10
min_am = 5
training_number = 10000
sample_careers, sample_labels = data_getter.get_sample_paths(
data_loc, samples_per_career, min_length, min_am)
training_careers = sample_careers[0:training_number]
training_labels = sample_labels[0:training_number]
testing_careers = sample_careers[training_number:]
testing_labels = sample_labels[training_number:]
#Null model first.
null_score = null_model_classifier(testing_careers, testing_labels)
print('null score: ', null_score)
#Now ML model.
model = svm.SVC(class_weight='balanced')
model.fit(training_careers, training_labels)
score = model.score(testing_careers, testing_labels)
print('model score: ', score)
def main():
data_loc = 'no_name_counts_actresses.txt'
samples_per_career = 5
min_length = 10
min_am = 5
training_number = 10000
sample_careers, sample_labels = data_getter.get_sample_paths(
data_loc, samples_per_career, min_length, min_am)
training_careers = sample_careers[0:training_number]
training_labels = sample_labels[0:training_number]
testing_careers = sample_careers[training_number:]
testing_labels = sample_labels[training_number:]
#Null model first.
null_score = null_model_classifier(testing_careers, testing_labels)
print('null score: ', null_score)
input_size = len(training_careers[0])
#Resize the input data if needed.
training_careers = training_careers.reshape(training_careers.shape + (1, ))
testing_careers = testing_careers.reshape(testing_careers.shape + (1, ))
print(training_careers.shape)
#Now ML model.
model = keras.models.Sequential()
#model.add(keras.layers.Dense(32, activation='relu', input_shape=(input_size,)))
#model.add(keras.layers.Dense(16, activation='relu'))
#model.add(keras.layers.Dense(16, activation='relu'))
#model.add(keras.layers.Dense(1, activation='sigmoid'))
model.add(
keras.layers.Conv1D(16,
5,
activation='relu',
input_shape=(None, training_careers.shape[-1])))
model.add(keras.layers.MaxPooling1D(5))
model.add(keras.layers.Conv1D(16, 5, activation='relu'))
model.add(keras.layers.GlobalMaxPooling1D())
model.add(keras.layers.Dense(16, activation='relu'))
model.add(keras.layers.Dense(16, activation='relu'))
model.add(keras.layers.Dense(1, activation='sigmoid'))
model.compile(optimizer='rmsprop',
loss='binary_crossentropy',
metrics=['acc'])
history = model.fit(training_careers,
training_labels,
epochs=20,
batch_size=256,
validation_data=(testing_careers, testing_labels))
#results_1 = model.evaluate(testing_careers, testing_labels)
history_dict = history.history
#print(history_dict.keys())
#Print section.
loss_values = history_dict['loss']
val_loss_values = history_dict['val_loss']
acc_values = history_dict['acc']
vall_acc_values = history_dict['val_acc']
epochs = range(1, len(acc_values) + 1)
#Plot the training values and validation loss.
plt.plot(epochs, loss_values, 'bo', label='Training loss')
plt.plot(epochs, val_loss_values, 'b', label='Validation loss')
plt.title('Training value and validation loss')
plt.xlabel('Epochs')
plt.ylabel('Loss')
plt.legend
plt.savefig('acting_class_training_and_validation_loss.png')
plt.close()
#Plot the training and validation accuracy.
plt.plot(epochs, acc_values, 'bo', label='Training acc')
plt.plot(epochs, vall_acc_values, 'b', label='Validation acc')
plt.title('Training and validation accuracy')
plt.xlabel('Epochs')
plt.ylabel('Loss')
plt.legend()
plt.savefig('acting_class_traing_and validation_acc.png')
plt.close()