verbose=True,
weightdecay=0.01)
trainer.trainUntilConvergence()
#trainer.trainEpochs(5)
print "trained"
#trainer.trainEpochs(5)
# Return a functor that wraps calling predict
return NeuralNetworkClassifier(trainer)
if __name__ == "__main__":
# First obtain our training and testing data
# Training has 50K samples, Testing 100K
Xt, Yt, Xv = load_validation_data()
# Run Neural Network over training data
classifier = classify(Xt, Yt)
# Prepare validation data and predict
tstdata = ClassificationDataSet(Xv.shape[1], 1, nb_classes=2)
tstdata.setField('input', Xv)
tstdata._convertToOneOfMany() # one output neuron per class
predictions = classifier.predict(tstdata)
# Write prediction to file
write_test_prediction("out_nn.txt", np.array(majority))
# build neural net and train it
net = buildNetwork(trndata.indim, n_hidden, trndata.outdim, outclass=SoftmaxLayer)
trainer = BackpropTrainer(net, dataset=trndata, momentum=0.1, verbose=True, weightdecay=0.01)
trainer.trainUntilConvergence()
#trainer.trainEpochs(5)
print "trained"
#trainer.trainEpochs(5)
# Return a functor that wraps calling predict
return NeuralNetworkClassifier(trainer)
if __name__ == "__main__":
# First obtain our training and testing data
# Training has 50K samples, Testing 100K
Xt, Yt, Xv = load_validation_data()
# Run Neural Network over training data
classifier = classify(Xt, Yt)
# Prepare validation data and predict
tstdata = ClassificationDataSet(Xv.shape[1], 1, nb_classes=2)
tstdata.setField('input', Xv)
tstdata._convertToOneOfMany() # one output neuron per class
predictions = classifier.predict(tstdata)
# Write prediction to file
write_test_prediction("out_nn.txt", np.array(majority))
adaboost.train(Xt, Yt),
extra_randomized_trees.train(Xt, Yt),
gradient_boost.train(Xt, Yt),
random_forest.train(Xt, Yt),
logistic_regression.train(Xt, Yt),
]
# Train another classifier on the ensembles output training predictions
# for each sample in the training data
training_predictions = np.mat(
[[c.predict(sample)[0] for c in classifiers] for sample in Xt])
meta_classifier = logistic_regression.train(training_predictions, Yt)
# Check results on training data
print "Accuracy for individual classifiers:", [
acc(Yt, c.predict(Xt)) for c in classifiers
]
predictions = np.mat([c.predict(Xt) for c in classifiers]).transpose()
print "Accuracy for ensemble classifier:", acc(
Yt, meta_classifier.predict(predictions))
### TEST DATA ###
# Predict on test data using the ensemble and meta classifier
predictions = np.mat([c.predict(Xv) for c in classifiers]).transpose()
final_predictions = meta_classifier.predict(predictions)
# Final accuracy
write_test_prediction("ensemble_predictions.txt",
np.array(final_predictions))
"""
gbc = GradientBoostingClassifier(verbose=1)
parameters = {'max_depth' : range(3,11),'n_estimators' : [400,500]}
classifier = GridSearchCV(gbc, parameters, scoring=metric)
classifier.fit(Xtrain, Ytrain)
return classifier
if __name__ == "__main__":
# Let's take our training data and train a decision tree
# on a subset. Scikit-learn provides a good module for cross-
# validation.
Xt, Xv, Yt, Yv = get_split_training_dataset()
Classifier = train(Xt, Yt)
print "Gradient Boost Classifier"
suite(Yv, Classifier.predict(Xv))
# smaller feature set
Xtimp, features = fclassify.get_important_data_features(Xt, Yt)
Xvimp = fclassify.compress_data_to_important_features(Xv, features)
ClassifierImp = train(Xtimp,Yt)
print "Gradient Boosts Classiifer, 25 important features"
suite(Yv, ClassifierImp.predict(Xvimp))
# save predictions on test data
X, Y, validation_data = load_validation_data()
predictions = Classifier.predict(validation_data)
filename = 'gradient_boost_predictions.txt'
write_test_prediction(filename, np.array(predictions))
def save_test_results(self, Xt, filename="nn_predictions.txt"):
predictions = self.predict(Xt)
write_test_prediction(filename, np.array(predictions))
#!/usr/bin/python2
# This is an optimized version of gradient boost.
import sys
import gradient_boost
import feature_selection_trees as fclassify
from util import write_test_prediction, load_validation_data
from metrics import acc
import numpy as np
if __name__ == "__main__":
# First obtain our training and testing data
Xt, Yt, Xv = load_validation_data()
# Train a gradietn boost classifier on it.
gboost = gradient_boost.train(Xt, Yt)
Yhat = gboost.predict(Xv)
# Final accuracy
write_test_prediction("gboost_optimal_2.txt", Yhat)
def save_test_results(self, Xt, filename="nn_predictions.txt"):
predictions = self.predict(Xt)
write_test_prediction(filename, np.array(predictions))
print "Accuracy for ensemble classifier:", acc(Yt2, meta_classifier.predict(predictions))
else:
# Now, we train each classifier on the training data
classifiers = [
adaboost.train(Xt, Yt),
extra_randomized_trees.train(Xt, Yt),
gradient_boost.train(Xt, Yt),
random_forest.train(Xt, Yt),
logistic_regression.train(Xt, Yt),
]
# Train another classifier on the ensembles output training predictions
# for each sample in the training data
training_predictions = np.mat([[c.predict(sample)[0] for c in classifiers] for sample in Xt])
meta_classifier = logistic_regression.train(training_predictions, Yt)
# Check results on training data
print "Accuracy for individual classifiers:", [acc(Yt, c.predict(Xt)) for c in classifiers]
predictions = np.mat([c.predict(Xt) for c in classifiers]).transpose()
print "Accuracy for ensemble classifier:", acc(Yt, meta_classifier.predict(predictions))
### TEST DATA ###
# Predict on test data using the ensemble and meta classifier
predictions = np.mat([c.predict(Xv) for c in classifiers]).transpose()
final_predictions = meta_classifier.predict(predictions)
# Final accuracy
write_test_prediction("ensemble_predictions.txt", np.array(final_predictions))
