def test_LabelSpreading_knn(*data):
x, y, unlabeled_indices = data
y_train = np.copy(y)
y_train[unlabeled_indices] = -1
fig = plt.figure()
ax = fig.add_subplot(1, 1, 1)
alphas = np.linspace(0.01, 1, num=10, endpoint=True)
Ks = [1, 2, 3, 4, 5, 8, 10, 15, 20, 25, 30, 35, 40, 50]
colors = (
(1, 0, 0), (0, 1, 0), (0, 0, 1), (0.5, 0.5, 0), (0, 0.5, 0.5), (0.5, 0, 0.5), (0.4, 0.6, 0), (0.6, 0.4, 0), \
(0, 0.6, 0.4), (0.5, 0.3, 0.2)) # 颜色集合，不同的曲线用不同的颜色
# 训练并绘图
for alpha, color in zip(alphas, colors):
scores = []
for K in Ks:
clf = LabelSpreading(max_iter=100, n_neighbors=K, alpha=alpha, kernel='knn')
clf.fit(x, y_train)
scores.append(clf.score(x[unlabeled_indices], y[unlabeled_indices]))
ax.plot(Ks, scores, label=r"$\alpha=%s$" % alpha, color=color)
# 设置图形
ax.set_xlabel(r"k")
ax.set_ylabel("score")
ax.legend(loc='best')
ax.set_title("LabelSpreading knn kernel")
plt.show()
def run_lp_bow_runtime_vocabulary(nbr, str_list, neighbors):
i = 0
avg_f1 = 0
avg_accuracy = 0
while i < 10:
dataset = Dataset(categories)
dataset.load_preprocessed(categories)
dataset.split_train_true(nbr)
print_v2_test_docs_vocabulary_labeled(categories)
dataset.load_preprocessed_test_vocabulary_labeled_in_use(categories)
vectorizer = CountVectorizer(vocabulary=Vocabulary.get_vocabulary(categories))
vectors = vectorizer.fit_transform(dataset.train['data'])
clf = LabelSpreading(kernel='knn', n_neighbors=neighbors).fit(vectors.todense(), dataset.train['target'])
test_vec = vectorizer.transform(dataset.test['data'])
pred = clf.predict(test_vec.todense())
avg_f1 += metrics.f1_score(dataset.test['target'], pred, average='macro')
avg_accuracy += clf.score(test_vec.todense(), dataset.test['target'])
i += 1
avg_accuracy = avg_accuracy/10
avg_f1 = avg_f1/10
str_list.extend(["KNN BOW runtime voc Avg f1: " + avg_f1.__str__(), "KNN BOW runtime vod Avg acc: "
+ avg_accuracy.__str__()])
print("Avg f1: " + avg_f1.__str__())
print("Avg acc: " + avg_accuracy.__str__())
def test_LabelSpreading_rbf(*data):
x, y, unlabeled_indices = data
y_train = np.copy(y)
y_train[unlabeled_indices] = -1
fig = plt.figure()
ax = fig.add_subplot(1, 1, 1)
alphas = np.linspace(0.01, 1, num=10, endpoint=True)
gammas = np.logspace(-2, 2, num=50)
colors = ((1, 0, 0), (0, 1, 0), (0, 0, 1), (0.5, 0.5, 0), (0, 0.5, 0.5), (0.5, 0, 0.5), (0.4, 0.6, 0), (0.6, 0.4, 0) \
, (0, 0.6, 0.4), (0.5, 0.3, 0.2)) # 颜色集合，不同的曲线用不同的颜色
# 训练并绘图
for alpha, color in zip(alphas, colors):
scores = []
for gamma in gammas:
clf = LabelSpreading(max_iter=100, gamma=gamma, alpha=alpha, kernel='rbf')
clf.fit(x, y_train)
scores.append(clf.score(x[unlabeled_indices], y[unlabeled_indices]))
ax.plot(gammas, scores, label=r"$\alpha=%s$" % alpha, color=color)
# 设置图形
ax.set_xlabel(r"$\gamma$")
ax.set_ylabel("score")
ax.set_xscale("log")
ax.legend(loc='best')
ax.set_title("LabelSpreading rbf kernel")
plt.show()
def test_LabelSpreading(*data):
X,y,unlabeled_indices = data
y_train = np.copy(y)
y_train[unlabeled_indices] = -1
clf = LabelSpreading(max_iter=1000, kernel='knn',gamma = 0.1)
clf.fit(X,y_train)
true_labels = y[unlabeled_indices]
predicted_labels = clf.transduction_[unlabeled_indices]
print('Accuracy : %f' %clf.score(X[unlabeled_indices],true_labels))
print('Accuracy : %f' %metrics.accuracy_score(true_labels,predicted_labels))
def test_LabelSpreading_alpha_gamma(*data):
X,y,unlabeled_indices = data
y_train = np.copy(y)
y_train[unlabeled_indices] = -1
alphas = np.logspace(-2,-1,num = 10)
gammas = np.logspace(-2,2,num = 10)
fig = plt.figure()
ax = fig.add_subplot(1,1,1)
for i,alpha in enumerate(alphas):
scores = []
for gamma in gammas:
clf = LabelSpreading(max_iter=1000, kernel='knn',gamma = gamma, alpha = alpha)
clf.fit(X,y_train)
true_labels = y[unlabeled_indices]
scores.append(clf.score(X[unlabeled_indices],true_labels))
ax.plot(alphas,scores,label='alpha = %f' %alpha)
ax.set_xscale('log')
ax.legend()
def run_lp_tfidf(nbr, str_list, neighbors):
i = 0
avg_f1 = 0
avg_accuracy = 0
while i < 10:
dataset = Dataset(categories)
dataset.split_train_true(nbr)
vectorizer = TfidfVectorizer()
vectors = vectorizer.fit_transform(dataset.train['data'])
clf = LabelSpreading(kernel='knn', n_neighbors=neighbors).fit(vectors.todense(), dataset.train['target'])
test_vec = vectorizer.transform(dataset.test['data'])
pred = clf.predict(test_vec.todense())
avg_f1 += metrics.f1_score(dataset.test['target'], pred, average='macro')
avg_accuracy += clf.score(test_vec.todense(), dataset.test['target'])
i += 1
avg_accuracy = avg_accuracy/10
avg_f1 = avg_f1/10
str_list.extend(["KNN TF-IDF Avg f1: " + avg_f1.__str__(), "KNN TF-IDF Avg acc: " + avg_accuracy.__str__()])
print("Avg f1: " + avg_f1.__str__())
print("Avg acc: " + avg_accuracy.__str__())
def label_spreading(x_train, y_train, x_test, y_test):
from sklearn.semi_supervised import LabelSpreading
sel = LabelSpreading()
sel.fit(x_train, y_train)
value = sel.score(x_test, y_test)
return "{0:.2f}".format(value)
idxs = np.random.choice(X_train.shape[0], replace = False, size=n_unlabeled)
y = np.asarray(Y_train)
for i in idxs:
y[i] = -1
Y_train = y
# Train model and print statistics (use 'knn' as kernel)
from sklearn.semi_supervised import LabelSpreading
model = LabelSpreading(kernel = 'knn', n_neighbors = 10, max_iter=1000).fit(X_train, Y_train)
print("Percentage of correct predictions = {}".format(round(100*model.score(X_test, Y_test),2)))
pred = model.predict(X_test) == Y_test
print("Correct: {}".format(np.count_nonzero(pred==True)),"/",
"Incorrect: {}".format(np.count_nonzero(pred==False)))
Z1 = model.predict(X_test).reshape(Y_test.size,1)
Z2 = np.asarray(Y_test).reshape(Y_test.size,1)
Z3 = np.around(model.predict_proba(X_test),decimals=2)
data = np.concatenate((Z1,Z2,Z3),axis=1)
outcome = pd.DataFrame(data, columns = ["Predicted Label",
"Actual Label",
"Prob. Label = 0.0",
"Prob. Label = 1.0"])
indicesToKeep = outcome["Predicted Label"] != outcome["Actual Label"]
print("False predictions with associated class probabilities:\n{}".format(outcome[indicesToKeep]))
hist, bins = np.histogram(
lables,
bins=[-0.1, 0.1, 1.1, 2.1, 3.1, 4.1, 5.1, 6.1, 7.1, 8.1, 9.1, 10.1])
print(hist)
print(bins)
print(train_labeled.shape)
print(train_labeled[:, 0])
train_unlabeled = sklearn.preprocessing.scale(train_unlabeled)
features = sklearn.preprocessing.scale(features)
lp = LabelSpreading(kernel='knn',
gamma=20,
n_neighbors=7,
alpha=0.2,
max_iter=50,
tol=0.01,
n_jobs=-1)
y = lables
for i in range(21000):
y = np.concatenate((y, np.array([-1])), axis=0)
all_data = np.concatenate((features, train_unlabeled), axis=0)
lp.fit(all_data, y)
Yresult = lp.predict(all_data)
print(lp.score(all_data, Yresult))
np.savetxt('semiLabelsOfUnlabeled2.csv', Yresult, delimiter=",")
gamma=p_gamma,
n_neighbors=p_neighbors,
alpha=p_alpha)
elif (p_ss_mod == 'LabSpr' and p_ss_kern == 'rbf'):
label_prop_model = LabelPropagation(kernel=p_ss_kern,
gamma=p_gamma,
n_neighbors=p_neighbors,
alpha=p_alpha,
max_iter=70)
else:
label_prop_model = dic_ss_mod[p_ss_mod](kernel=p_ss_kern,
gamma=p_gamma,
n_neighbors=p_neighbors)
print('Start to fit. Run for shelter!')
label_prop_model.fit(X_tot, y_tot)
temp_acc = label_prop_model.score(X_valid_lab, y_valid)
print('{} / {} :accuracy = {}'.format(i, p_manyfit, temp_acc))
RESULT_ACC_SS += temp_acc
y_tot = label_prop_model.transduction_
y_submit = label_prop_model.predict(X_submit)
save_to_csv(X_tot, y_tot, X_valid_lab, y_valid)
RESULT_ACC_SS /= p_manyfit
json_dict['ss_accuracy'] = RESULT_ACC_SS
print('accuracy obtained on the test set of the ss algo:', RESULT_ACC_SS)
else:
init_variables()
#PCA preprocessing
if (PCA_MODE):
pca_preprocess()
X_tot, y_tot, X_valid, y_valid = load_xy()
def LabelData(LabelX,
unLabelX,
Labely,
unLabely,
testX,
testy,
batch_id=0,
save_model=False):
LabelXLen = LabelX.shape[0]
print("LabeledCellNames", LabelX)
X = pd.concat([LabelX, unLabelX], axis=0, join='inner')
y = np.append(Labely, unLabely)
Features = X.columns.values.tolist()
testX = testX.loc[:, Features]
# Knn LabelSpreading
label_spread = LabelSpreading(kernel='knn', alpha=0.8, max_iter=5)
print("==== X ====")
print(X)
print("==== Y ====")
print(y)
label_spread.fit(X, y)
output_labels = label_spread.transduction_
score = label_spread.score(testX, testy)
output_labels = le.inverse_transform(output_labels)
CellNames = X.index.values.tolist()
CellResult = {
"CellName": CellNames[LabelXLen + 1:],
"CellType": output_labels[LabelXLen + 1:]
}
# Result = pd.DataFrame(data=CellResult)
# Result.to_csv("./result/%d.csv"%(batch_id), columns=['CellName','CellType'], index=False)
# accuracy
print("score : ", score)
PredictY = label_spread.predict(testX)
PredictYLabels = le.inverse_transform(PredictY)
TrueYLabels = le.inverse_transform(testy)
PredictResult = {"trueLabel": TrueYLabels, "predictLabel": PredictYLabels}
# PredictResult = pd.DataFrame(data=PredictResult)
# PredictResult.to_csv("./result/predict_result_%d.csv"%(batch_id), columns=['trueLabel','predictLabel'], index=False)
# Label Distribution
print("======= label_spread.label_distributions_ =======")
print(label_spread.label_distributions_)
LabelXIndexs = LabelX.index
indexs = X.index
ClassLabels = le.inverse_transform(label_spread.classes_)
print(ClassLabels)
LabelDistribution = pd.DataFrame(data=label_spread.label_distributions_,
index=indexs,
columns=ClassLabels)
LabelDistribution = LabelDistribution.drop(index=LabelXIndexs)
# LabelDistribution.to_csv("./result/test/LabelDistribution.csv")
if save_model:
with open('./result/%s/clf.pickle' % (FileName), 'wb') as f:
pickle.dump(label_spread, f)
return CellResult, LabelDistribution
# classification
# use max_iter=10 when 20 categories
clf_rbf = LabelPropagation(kernel='rbf', gamma=5).fit(vectors_rbf.todense(), dataset_rbf.train['target'])
clf_knn = LabelSpreading(kernel='knn', n_neighbors=10).fit(vectors_knn.todense(), dataset_knn.train['target'])
test_vec_rbf = vectorizer_rbf.transform(dataset_rbf.test['data'])
test_vec_knn = vectorizer_knn.transform(dataset_knn.test['data'])
print('----PREDICTIONS----')
pred_rbf = clf_rbf.predict(test_vec_rbf.todense())
pred_knn = clf_knn.predict(test_vec_knn.todense())
print('f1 score rbf: ', metrics.f1_score(dataset_rbf.test['target'], pred_rbf, average='macro'))
print('clf score rbf: ', clf_rbf.score(test_vec_rbf.todense(), dataset_rbf.test['target']))
print('f1 score knn: ', metrics.f1_score(dataset_knn.test['target'], pred_knn, average='macro'))
print('clf score knn: ', clf_knn.score(test_vec_knn.todense(), dataset_knn.test['target']))
np.set_printoptions(precision=2)
""""
# Plot non-normalized confusion matrix
plot_confusion_matrix(dataset_rbf.test['target'], pred_rbf, classes=categories,
title='Confusion matrix (RBF), without normalization')
# Plot normalized confusion matrix
plot_confusion_matrix(dataset_rbf.test['target'], pred_rbf, classes=categories, normalize=True,
title='Normalized confusion matrix (RBF)')
plt.show()
# Plot non-normalized confusion matrix
#print(gridsearch.best_params_)
print('got Vectors')
model = LabelSpreading(kernel='rbf')
params = {'gamma': [0.1, 1.0, 10.0, 30.0, 50.0, 80.0, 100.0, 300.0], 'max_iter': [10, 100, 1000], 'alpha': [0.2, 0.4, 0.6, 0.8]}
scoreDict = {}
for max_iter in params['max_iter']:
model.max_iter = max_iter
for alpha in params['alpha']:
model.alpha = alpha
for gamma in params['gamma']:
model.gamma = gamma
model.fit(list(data), list(unlab))
score = model.score(list(testData), list(testLabels))
print(score, ' gamma = ', gamma, ' max_iter = ', max_iter, ' alpha = ', alpha)
if (score in scoreDict):
scoreDict[score].append(
'gamma = ' + str(gamma) + ' max_iter = ' + str(max_iter) + ' alpha = ' + str(alpha))
else:
scoreDict[score] = [
'gamma = ' + str(gamma) + ' max_iter = ' + str(max_iter) + ' alpha = ' + str(alpha)]
knnModel = LabelSpreading(kernel='knn')
knnParams = {'n_neighbors': [1, 4, 9, 16], 'max_iter': [10, 100, 1000], 'alpha': [0.2, 0.4, 0.6, 0.8]}
for max_iter in knnParams['max_iter']:
model.max_iter = max_iter
for alpha in knnParams['alpha']:
model.alpha = alpha
for n_neighbors in knnParams['n_neighbors']:
def run_methods(x_c, y, x_e, z_c, z_y, z_e):
x = np.concatenate((x_c, x_e), axis=1)
z = np.concatenate((z_c, z_e), axis=1)
# Baseline: Linear Logistic Regression
lin_lr = LogisticRegression(random_state=0,
solver='liblinear').fit(x, y.ravel())
acc_lin_lr = lin_lr.score(z, z_y)
# hard_label_lin_lr = lin_lr.predict(z)
# soft_label_lin_lr = lin_lr.predict_proba(z)[:, 1]
# TRANSDUCTIVE APPROACHES
# merge labelled and unlabelled data (with label -1) for transductive methods
x_merged = np.concatenate((x, z))
y_merged = np.concatenate((y, -1 * np.ones(
(z.shape[0], 1)))).ravel().astype(int)
# Baseline: Linear TSVM: https://github.com/tmadl/semisup-learn/tree/master/methods
lin_tsvm = SKTSVM(kernel='linear')
lin_tsvm.fit(x_merged, y_merged)
acc_lin_tsvm = lin_tsvm.score(z, z_y)
# hard_label_lin_tsvm = lin_tsvm.predict(z)
# soft_label_lin_tsvm = lin_tsvm.predict_proba(z)[:, 1]
# Baseline: Non-Linear TSVM: https://github.com/tmadl/semisup-learn/tree/master/methods
rbf_tsvm = SKTSVM(kernel='RBF')
rbf_tsvm.fit(x_merged, y_merged)
acc_rbf_tsvm = rbf_tsvm.score(z, z_y)
# hard_label_rbf_tsvm = rbf_tsvm.predict(z)
# soft_label_rbf_tsvm = rbf_tsvm.predict_proba(z)[:, 1]
# Baseline: Label Propagation RBF weights
try:
rbf_label_prop = LabelPropagation(kernel='rbf')
rbf_label_prop.fit(x_merged, y_merged)
acc_rbf_label_prop = rbf_label_prop.score(z, z_y)
# hard_label_rbf_label_prop= rbf_label_prop.predict(z)
# soft_label_rbf_label_prop = rbf_label_prop.predict_proba(z)[:, 1]
except:
acc_rbf_label_prop = []
print 'rbf label prop did not work'
# Baseline: Label Spreading with RBF weights
try:
rbf_label_spread = LabelSpreading(kernel='rbf')
rbf_label_spread.fit(x_merged, y_merged)
acc_rbf_label_spread = rbf_label_spread.score(z, z_y)
# hard_label_rbf_label_spread = rbf_label_spread.predict(z)
# soft_label_rbf_label_spread = rbf_label_spread.predict_proba(z)[:, 1]
except:
acc_rbf_label_spread = []
print 'rbf label spread did not work '
# THE K-NN VERSIONS ARE UNSTABLE UNLESS USING LARGE K
# Baseline: Label Propagation with k-NN weights
try:
knn_label_prop = LabelPropagation(kernel='knn', n_neighbors=11)
knn_label_prop.fit(x_merged, y_merged)
acc_knn_label_prop = knn_label_prop.score(z, z_y)
# hard_label_knn_label_prop = knn_label_prop.predict(z)
# soft_label_knn_label_prop = knn_label_prop.predict_proba(z)[:, 1]
except:
acc_knn_label_prop = []
print 'knn label prop did not work'
# Baseline: Label Spreading with k-NN weights
try:
knn_label_spread = LabelSpreading(kernel='knn', n_neighbors=11)
knn_label_spread.fit(x_merged, y_merged)
acc_knn_label_spread = knn_label_spread.score(z, z_y)
# hard_label_knn_label_spread = knn_label_spread.predict(z)
# soft_label_knn_label_spread = knn_label_spread.predict_proba(z)[:, 1]
except:
acc_knn_label_spread = []
print 'knn label spread did not work'
# Generative Models
# Semi-generative model on labelled data only
a_y, b_y, a_e0, a_e1, b_0, b_1, cov_e0, cov_e1 = soft_label_EM(
x_c, y, x_e, z_c, z_e, converged=True)
soft_label_semigen = predict_class_probs(z_c, z_e, a_y, b_y, a_e0, a_e1,
b_0, b_1, cov_e0, cov_e1)
hard_label_semigen = soft_label_semigen > 0.5
acc_semigen_labelled = np.mean(hard_label_semigen == z_y)
# EM with soft labels
a_y, b_y, a_e0, a_e1, b_0, b_1, cov_e0, cov_e1 = soft_label_EM(
x_c, y, x_e, z_c, z_e)
soft_label_soft_EM = predict_class_probs(z_c, z_e, a_y, b_y, a_e0, a_e1,
b_0, b_1, cov_e0, cov_e1)
hard_label_soft_EM = soft_label_soft_EM > 0.5
acc_soft_EM = np.mean(hard_label_soft_EM == z_y)
# EM with hard labels
a_y, b_y, a_e0, a_e1, b_0, b_1, cov_e0, cov_e1 = hard_label_EM(
x_c, y, x_e, z_c, z_e)
soft_label_hard_EM = predict_class_probs(z_c, z_e, a_y, b_y, a_e0, a_e1,
b_0, b_1, cov_e0, cov_e1)
hard_label_hard_EM = soft_label_hard_EM > 0.5
acc_hard_EM = np.mean(hard_label_hard_EM == z_y)
# Conditional label prop
acc_cond_prop = conditional_prop(x_c, y, x_e, z_c, z_y, z_e)
return acc_lin_lr, acc_lin_tsvm, acc_rbf_tsvm, acc_rbf_label_prop, acc_rbf_label_spread, acc_knn_label_prop,\
acc_knn_label_spread, acc_semigen_labelled, acc_soft_EM, acc_hard_EM, acc_cond_prop
# In[356]:
# Test Label Spreading by cross validation
skf = StratifiedKFold(n_splits=5)
score0 = []
score0_holdout = []
score1 = []
score1_holdout = []
for i_train, i_test in skf.split(X_train2_, Y_train2_.argmax(axis=1)):
X_train3, y_train3 = X_train2_[i_train], Y_train2_[i_train].argmax(axis=1)
X_holdout3, y_holdout3 = X_train2_[i_test], Y_train2_[i_test].argmax(
axis=1)
n_holdout3 = len(y_holdout3)
ls0 = LabelSpreading(kernel='rbf', gamma=2, n_neighbors=4)
ls0.fit(X_train3, y_train3)
score0.append(ls0.score(X_holdout3, y_holdout3))
score0_holdout.append(ls0.score(X_holdout2, Y_holdout2.argmax(axis=1)))
print(' Supervised score: {:.4f} (holdout {:.4f})'.format(
score0[-1], score0_holdout[-1]))
ls1 = LabelSpreading(kernel='rbf', gamma=2, n_neighbors=4)
ls1.fit(np.vstack((X_train3, X_holdout3)),
np.concatenate((y_train3, np.full(n_holdout3, -1))))
score1.append(ls1.score(X_holdout3, y_holdout3))
score1_holdout.append(ls1.score(X_holdout2, Y_holdout2.argmax(axis=1)))
print(' Semi-Supervised score: {:.4f} (holdout {:.4f})'.format(
score1[-1], score1_holdout[-1]))
print('Mean supervised: {:.4f} (holdout {:.4f})'.format(
np.mean(score0), np.mean(score0_holdout)))
print('Mean semi-supervised: {:.4f} (holdout {:.4f})'.format(
np.mean(score1), np.mean(score1_holdout)))
def label_spreading(self,
kernel='rbf',
gamma=20,
n_neighbors=7,
alpha=0.2,
max_iter=30,
tol=0.001,
n_jobs=1):
"""
LabelSpreading model for semi-supervised learning
This model is similar to the basic Label Propagation algorithm,
but uses affinity matrix based on the normalized graph Laplacian
and soft clamping across the labels.
Parameters
----------
kernel : {'knn', 'rbf', callable}
String identifier for kernel function to use or the kernel function
itself. Only 'rbf' and 'knn' strings are valid inputs. The function
passed should take two inputs, each of shape [n_samples, n_features],
and return a [n_samples, n_samples] shaped weight matrix
gamma : float
parameter for rbf kernel
n_neighbors : integer > 0
parameter for knn kernel
alpha : float
Clamping factor. A value in [0, 1] that specifies the relative amount
that an instance should adopt the information from its neighbors as
opposed to its initial label.
alpha=0 means keeping the initial label information; alpha=1 means
replacing all initial information.
max_iter : integer
maximum number of iterations allowed
tol : float
Convergence tolerance: threshold to consider the system at steady
state
n_jobs : int or None, optional (default=None)
The number of parallel jobs to run.
``None`` means 1 unless in a :obj:`joblib.parallel_backend` context.
``-1`` means using all processors. See :term:`Glossary <n_jobs>`
for more details.
Returns
-------
score : the score of learning model on test data
Example
--------
>>> labeled_path = "../data/labeled.csv"
>>> unlabeled_path = "../data/unlabeled.csv"
>>> mtl = MultiTaskLearner(labeled_path, unlabeled_path)
>>> encoding = mtl.embed(word_length=5)
>>> X, y, X_t, y_t = train_test_split(mtl.sequences, mtl.labels, test_size=0.33)
>>> score = mtl.semi_supervised_learner(X, y, X_t, y_t, ssl="label_spreading")
"""
model = LabelSpreading(kernel=kernel,
gamma=gamma,
n_neighbors=n_neighbors,
alpha=alpha,
max_iter=max_iter,
tol=tol,
n_jobs=n_jobs)
model.fit(self.X, self.y)
return model.score(self.X_t, self.y_t)
pred = model4.predict(X_test) == Y_test
statistics.loc[3] = ["SS Naive Bayes",
round(100*model4.score(X_test, Y_test),2),
np.count_nonzero(pred==True),
np.count_nonzero(pred==False),
round(100*idxs.size/Y_train.size,2),
training_time4]
# Train semi-supervised LabelSpreading model, predict (use 'knn' as kernel) and collect statistics
from sklearn.semi_supervised import LabelSpreading
start_time5 = time.time_ns()
model5 = LabelSpreading(kernel = 'knn', n_neighbors = 10, max_iter=1000).fit(X_train, Y_train)
training_time5 = time.time_ns() - start_time5
pred = model5.predict(X_test) == Y_test
statistics.loc[4] = ["SS Label Spreading",
round(100*model5.score(X_test, Y_test),2),
np.count_nonzero(pred==True),
np.count_nonzero(pred==False),
round(100*idxs.size/Y_train.size,2),
training_time5]
# Print summary statistics
print(statistics)
max_iter=1000).fit(vectors.todense(),
dataset.train['target'])
clf_knn = LabelSpreading(kernel='knn', n_neighbors=5,
max_iter=1000).fit(vectors_knn.todense(),
dataset_knn.train['target'])
test_vec_rbf = vectorizer_rbf.transform(dataset.test['data'])
test_vec_knn = vectorizer_knn.transform(dataset_knn.test['data'])
print('----PREDICTIONS----')
pred_rbf = clf_rbf.predict(test_vec_rbf.todense())
pred_knn = clf_knn.predict(test_vec_knn.todense())
print('f1 score rbf: ',
metrics.f1_score(dataset.test['target'], pred_rbf, average='macro'))
print('clf score rbf: ',
clf_rbf.score(test_vec_rbf.todense(), dataset.test['target']))
print('f1 score knn: ',
metrics.f1_score(dataset_knn.test['target'], pred_knn, average='macro'))
print('clf score knn: ',
clf_knn.score(test_vec_knn.todense(), dataset_knn.test['target']))
np.set_printoptions(precision=2)
"""
# Plot non-normalized confusion matrix
plot_confusion_matrix(dataset.test['target'], pred_rbf, classes=categories,
title='Confusion matrix (RBF with vocabulary), without normalization')
# Plot normalized confusion matrix
plot_confusion_matrix(dataset.test['target'], pred_rbf, classes=categories, normalize=True,
title='Normalized confusion matrix (RBF with vocabulary)')
plt.show()
