def main(): # read hard and easy
estimator = misvm.MISVM(kernel=kernel,
gamma=gamma,
C=C,
verbose=True,
max_iters=100)
mil.train(estimator=estimator, resultSuperDirPath=path)
def cross_validation():
estimator = misvm.MISVM(kernel=kernel,
gamma=gamma,
C=C,
verbose=True,
max_iters=200)
mil.crossValidation(estimator, 5, path)
def leave_one_person_out(n_jobs=8):
estimator = misvm.MISVM(kernel=kernel,
gamma=gamma,
C=C,
verbose=True,
max_iters=200)
mil.leaveOnePersonOut(estimator=estimator,
resultSuperDirPath=path,
n_jobs=n_jobs,
trainAcc=True)
def misvm_classify(verbose, output, vectors, labels):
# perform the actual misvm classification
if verbose:
print "Creating train and test bags and labels..."
bags = [numpy.array(vectors[v], dtype=float)
for v in vectors] # numpy-format matrix for use in misvm
labels = numpy.array([labels[l] for l in labels],
dtype=float) # numpy-format labels for use in misvm
# Spilt dataset into train and test sets
train_bags = []
train_labels = []
test_bags = []
test_labels = []
for i in range(len(labels)):
if i % 2 == 0:
train_bags.append(bags[i])
train_labels.append(labels[i])
else:
test_bags.append(bags[i])
test_labels.append(labels[i])
if verbose:
print "MISVM Classifying..."
if output != 'NONE':
sys.stdout = open(output, 'w')
# establish classifiers
classifiers = {
'sbMIL': misvm.sbMIL(kernel='rbf', eta=0.1, C=1.0),
'SIL': misvm.SIL(kernel='rbf', C=1.0),
'MISVM': misvm.MISVM(kernel='rbf', C=1.0, max_iters=100),
}
# Train/Evaluate classifiers
accuracies = {}
for algorithm, classifier in classifiers.items():
classifier.fit(train_bags, train_labels)
predictions = classifier.predict(test_bags)
accuracies[algorithm] = numpy.average(
test_labels == numpy.sign(predictions))
for algorithm, accuracy in accuracies.items():
print '\n%s Accuracy: %.1f%%' % (algorithm, 100 * accuracy)
if output != 'NONE':
sys.stdout = sys.__stdout__ # reset stdout to normal
def __init__(self, cfg, **kwargs):
super(MISVMDetector, self).__init__(cfg, **kwargs)
self.classifier = misvm.MISVM(kernel='linear', C=1.0, max_iters=50)
def main(argv=sys.argv):
pos_train_file, neg_train_file, method, model_file, scale_file, score_file, res_file = process_options(
argv)
######train feature extraction
feature_matrix = []
for line in pos_train_file:
feature_vector = []
sequence_infor = line.split(',')
header = sequence_infor[0]
bag = sequence_infor[1]
sequence = sequence_infor[3].strip()
feature_vector.append(header)
feature_vector.append(bag)
feature_vector.extend(
kmer(sequence) + ksnpf(sequence) + nucleic_shift(sequence))
#feature_vector.extend(ksnpf(sequence))
feature_vector.append('1')
feature_matrix.append(feature_vector)
pos_train_file.close()
for line in neg_train_file:
feature_vector = []
sequence_infor = line.split(',')
header = sequence_infor[0]
bag = sequence_infor[1]
sequence = sequence_infor[3].strip()
feature_vector.append(header)
feature_vector.append(bag)
feature_vector.extend(
kmer(sequence) + ksnpf(sequence) + nucleic_shift(sequence))
#feature_vector.extend(ksnpf(sequence))
feature_vector.append('-1')
feature_matrix.append(feature_vector)
feature_array = np.array([b[2:-1] for b in feature_matrix],
dtype=np.float32)
min_max_scaler = preprocessing.MinMaxScaler(copy=True,
feature_range=(-1, 1))
feature_scaled = min_max_scaler.fit_transform(feature_array)
feature_matrix_T = map(list, zip(*feature_matrix))
feature_scaled_T = map(list, zip(*feature_scaled))
k = 0
train_feature_matrix_T = []
train_feature_matrix_T.append(feature_matrix_T[0])
train_feature_matrix_T.append(feature_matrix_T[1])
for i in range(len(feature_scaled_T)):
train_feature_matrix_T.append(feature_scaled_T[k])
k = k + 1
train_feature_matrix_T.append(feature_matrix_T[-1])
train_feature_matrix = map(list, zip(*train_feature_matrix_T))
neg_train_file.close()
np.savetxt("train_features.txt",
train_feature_matrix,
fmt='%s',
delimiter=',')
######put samples into bags
train_file_path = './train_features.txt'
[
train_bag_targets, train_bag_samples, train_bag_instance_targets,
sample_info
] = file_to_features(train_file_path)
if method == 0:
svc = misvm.miSVM(kernel='quadratic', C=5.4, max_iters=10)
elif method == 1:
svc = misvm.MISVM(kernel='quadratic', C=5.4, max_iters=10)
svc.fit(train_bag_samples, train_bag_targets)
#joblib.dump(svc,'./svc.pkl')
joblib.dump(svc, model_file)
joblib.dump(min_max_scaler, scale_file)
bag_predictions, inst_predictions = svc.predict(train_bag_samples, 1)
#score_file=open('training_score.txt','w+')
score_file.write("name\tscore\n")
for i in range(0, len(sample_info)):
score_file.write(sample_info[i] + "\t" + str(inst_predictions[i]) +
"\n")
score_file.close()
#reserved_samples=open('reserved_samples.txt','w+')
res_file.write("name\n")
for j in range(0, len(inst_predictions)):
if inst_predictions[j] >= 0:
res_file.write(sample_info[j] + "\n")
res_file.close()
def gridsearch(params_grid, cv=2):
estimator = misvm.MISVM(max_iters=250)
mil.gridSearch(estimator=estimator, params_grid=params_grid, cv=cv)