def main():
# In this example, we have three types of samples: class 0, 1, or 2. That
# is, each of our sample vectors falls into one of three classes. To keep
# this example very simple, each sample vector is zero everywhere except at
# one place. The non-zero dimension of each vector determines the class of
# the vector. So for example, the first element of samples has a class of 1
# because samples[0][1] is the only non-zero element of samples[0].
samples = [[0, 2, 0], [1, 0, 0], [0, 4, 0], [0, 0, 3]]
# Since we want to use a machine learning method to learn a 3-class
# classifier we need to record the labels of our samples. Here samples[i]
# has a class label of labels[i].
labels = [1, 0, 1, 2]
# Now that we have some training data we can tell the structural SVM to
# learn the parameters of our 3-class classifier model. The details of this
# will be explained later. For now, just note that it finds the weights
# (i.e. a vector of real valued parameters) such that predict_label(weights,
# sample) always returns the correct label for a sample vector.
problem = ThreeClassClassifierProblem(samples, labels)
weights = dlib.solve_structural_svm_problem(problem)
# Print the weights and then evaluate predict_label() on each of our
# training samples. Note that the correct label is predicted for each
# sample.
print(weights)
for k, s in enumerate(samples):
print("Predicted label for sample[{0}]: {1}".format(
k, predict_label(weights, s)))
def main():
# In this example, we have three types of samples: class 0, 1, or 2. That
# is, each of our sample vectors falls into one of three classes. To keep
# this example very simple, each sample vector is zero everywhere except at
# one place. The non-zero dimension of each vector determines the class of
# the vector. So for example, the first element of samples has a class of 1
# because samples[0][1] is the only non-zero element of samples[0].
samples = [[0, 2, 0], [1, 0, 0], [0, 4, 0], [0, 0, 3]]
# Since we want to use a machine learning method to learn a 3-class
# classifier we need to record the labels of our samples. Here samples[i]
# has a class label of labels[i].
labels = [1, 0, 1, 2]
# Now that we have some training data we can tell the structural SVM to
# learn the parameters of our 3-class classifier model. The details of this
# will be explained later. For now, just note that it finds the weights
# (i.e. a vector of real valued parameters) such that predict_label(weights,
# sample) always returns the correct label for a sample vector.
problem = ThreeClassClassifierProblem(samples, labels)
weights = dlib.solve_structural_svm_problem(problem)
# Print the weights and then evaluate predict_label() on each of our
# training samples. Note that the correct label is predicted for each
# sample.
print(weights)
for k, s in enumerate(samples):
print("Predicted label for sample[{0}]: {1}".format(
k, predict_label(weights, s)))
def main():
# In this example, we have three types of samples: class 0, 1, or 2. That is, each of
# our sample vectors falls into one of three classes. To keep this example very
# simple, each sample vector is zero everywhere except at one place. The non-zero
# dimension of each vector determines the class of the vector. So for example, the
# first element of samples has a class of 1 because samples[0][1] is the only non-zero
# element of samples[0].
samples = [[0, 2, 0], [1, 0, 0], [0, 4, 0], [0, 0, 3]]
# Since we want to use a machine learning method to learn a 3-class classifier we need
# to record the labels of our samples. Here samples[i] has a class label of labels[i].
labels = [1, 0, 1, 2]
# Now that we have some training data we can tell the structural SVM to learn the
# parameters of our 3-class classifier model. The details of this will be explained
# later. For now, just note that it finds the weights (i.e. a vector of real valued
# parameters) such that predict_label(weights, sample) always returns the correct label
# for a sample vector.
problem = three_class_classifier_problem(samples, labels)
weights = dlib.solve_structural_svm_problem(problem)
# Print the weights and then evaluate predict_label() on each of our training samples.
# Note that the correct label is predicted for each sample.
print weights
for i in range(len(samples)):
print "predicted label for sample[{0}]: {1}".format(
i, predict_label(weights, samples[i]))
def main():
# In this example, we have three types of samples: class 0, 1, or 2. That is, each of
# our sample vectors falls into one of three classes. To keep this example very
# simple, each sample vector is zero everywhere except at one place. The non-zero
# dimension of each vector determines the class of the vector. So for example, the
# first element of samples has a class of 1 because samples[0][1] is the only non-zero
# element of samples[0].
samples = [[0,2,0], [1,0,0], [0,4,0], [0,0,3]];
# Since we want to use a machine learning method to learn a 3-class classifier we need
# to record the labels of our samples. Here samples[i] has a class label of labels[i].
labels = [1,0,1,2]
# Now that we have some training data we can tell the structural SVM to learn the
# parameters of our 3-class classifier model. The details of this will be explained
# later. For now, just note that it finds the weights (i.e. a vector of real valued
# parameters) such that predict_label(weights, sample) always returns the correct label
# for a sample vector.
problem = three_class_classifier_problem(samples, labels)
weights = dlib.solve_structural_svm_problem(problem)
# Print the weights and then evaluate predict_label() on each of our training samples.
# Note that the correct label is predicted for each sample.
print weights
for i in range(len(samples)):
print "predicted label for sample[{0}]: {1}".format(i, predict_label(weights, samples[i]))
def main():
# for simple testing
# I recommend try training_sample['feature'][:30] and testing_sample['feature'][:30]
# and training_sample['labels'][:30] and testing_sample['labels'][:30]
# should give a high training accuracy and a low testing accuracy
# need to read 4001 + 1001 = 5002
dataset1 = read_OCR('OCRdataset/letter.data', 5002, 128)
(training_sample, testing_sample) = recompose_data(dataset1)
print "total training samples: ", len(training_sample['words'])
print "total testing samples: ", len(testing_sample['words'])
# problem = MultiClassClassifierProblem(training_sample['feature'][:30],training_sample['labels'][:30])
problem = MultiClassClassifierProblem(training_sample['feature'],
training_sample['labels'])
weights = dlib.solve_structural_svm_problem(problem)
# get training accuracy
predictions = []
# for samp in training_sample['feature'][:30]:
for samp in training_sample['feature']:
prediction = [0] * window_size
Nither = 4
max1 = 0
for k in range(Nither):
for iL in range(window_size):
for i in range(27):
temp_label = list(prediction)
temp_label[iL] = i
psi1 = problem.make_psi(samp, temp_label)
score1 = dlib.dot(weights, psi1)
if max1 < score1:
max1 = score1
prediction[iL] = i
predictions.append(prediction)
# print("weights", weights)
print predictions
# print "training accuracy=", accuracy_score(predictions, training_sample['labels'][:30])
print "training accuracy=", accuracy_score(predictions,
training_sample['labels'])
# get testing accuracy
te_predictions = []
# for samp in testing_sample['feature'][:30]:
for samp in testing_sample['feature']:
te_prediction = [0] * window_size
Nither = 4
max1 = 0
for k in range(Nither):
for iL in range(window_size):
for i in range(27):
temp_label = list(te_prediction)
temp_label[iL] = i
psi1 = problem.make_psi(samp, temp_label)
score1 = dlib.dot(weights, psi1)
if max1 < score1:
max1 = score1
te_prediction[iL] = i
te_predictions.append(te_prediction)
# print te_labels
# print te_predictions
# print "test accuracy=", accuracy_score(te_predictions, testing_sample['labels'][:30])
print "test accuracy=", accuracy_score(te_predictions,
testing_sample['labels'])
