def train_custom_one_vs_all(X_train, X_test, Y_train, topk):
#convert matrix to row for efficient splicing
Y_train = Y_train.tocsc()
tag_classifiers = []
num_training, numclasses = Y_train.shape
num_test_examples = X_test.shape[0]
# hold a vector mxk, containing top k prediction classes for each example, maintain m heaps for that
num_examples = X_test.shape[0]
num_classes = len(tag_classifiers)
topk_class_distances = []
for i in xrange(num_examples):
heap = []
topk_class_distances += [heap]
for j in xrange(numclasses):
# train on each class label for all the training examples
y = numpy.ravel(Y_train.getcol(j).todense())
clf = LogisticRegression(penalty='l2',
dual=False,
tol=0.0001,
C=0.8,
fit_intercept=True,
intercept_scaling=1)
clf.fit(X_train, y)
print "Trained for class", j
# get the decision for all test examples
decision = clf.densify().decision_function(X_test)
# for each test example add its decision value to the heap of top k decision values
for i in xrange(num_test_examples):
h = topk_class_distances[i]
if len(h) < topk: heapq.heappush(h, (decision[i], j))
else: heapq.heappushpop(h, (decision[i], j))
print "Predicted for class", j
#clean the decision values and store the class labels
class_label_indices = []
for i in xrange(num_examples):
topk_labels = [label for dist, label in topk_class_distances[i]]
class_label_indices += [topk_labels]
return class_label_indices
def train_custom_one_vs_all(X_train,X_test,Y_train,topk):
#convert matrix to row for efficient splicing
Y_train = Y_train.tocsc()
tag_classifiers = []
num_training,numclasses = Y_train.shape
num_test_examples = X_test.shape[0]
# hold a vector mxk, containing top k prediction classes for each example, maintain m heaps for that
num_examples = X_test.shape[0]
num_classes = len(tag_classifiers)
topk_class_distances = []
for i in xrange(num_examples):
heap = []
topk_class_distances += [heap]
for j in xrange(numclasses):
# train on each class label for all the training examples
y = numpy.ravel(Y_train.getcol(j).todense());
clf = LogisticRegression(penalty='l2',dual=False,tol=0.0001,C=0.8,fit_intercept=True,intercept_scaling=1)
clf.fit(X_train,y);
print "Trained for class",j
# get the decision for all test examples
decision = clf.densify().decision_function(X_test)
# for each test example add its decision value to the heap of top k decision values
for i in xrange(num_test_examples):
h = topk_class_distances[i]
if len(h) < topk: heapq.heappush(h,(decision[i],j))
else: heapq.heappushpop(h,(decision[i],j))
print "Predicted for class",j
#clean the decision values and store the class labels
class_label_indices = []
for i in xrange(num_examples):
topk_labels = [label for dist,label in topk_class_distances[i]]
class_label_indices += [topk_labels]
return class_label_indices
clf = RandomForestClassifier(n_estimators=100, max_depth=2,
random_state=0)
clf.fit(X_train, y_train)
print("Accuracy ",clf.score(X_test, y_test))
# Model Building
#using logistic regression
print("TRAINING PHASE")
logit = LogisticRegression()
logit.fit(X_train, y_train)
print("Accuracy ",logit.score(X_test, y_test))
print("coefficient :\n",logit.coef_)
print("Intercept:\n",logit.intercept_)
print(logit.densify())
print(logit.sparsify())
url_list = list(url_list)
print("TESTING PHASE")
X_predict = ["8.8.8.8"]
with open('logit.pickle', 'wb') as handle:
pickle.dump(logit, handle, protocol=2)
with open('vectorizer.pickle', 'wb') as handle:
pickle.dump(vectorizer, handle, protocol=2)
solver='liblinear', # for use with small datasets
multi_class='ovr') # stating this is a binary problem)
# training the model
clf.fit(x_train, y_train)
# attributes
classes = clf.classes_ # list of class labels
coeff = clf.coef_ # coefficients of the model
intercept = clf.intercept_ # the intercept of the model
n_iter = clf.n_iter_ # the number of iterations for each class - in the binary case it only returns one value
# now having a look at the methods
dec_func = clf.decision_function(
x_test) # the confidence score for each test data
density = clf.densify() # returns the coeffient matrix in densy array format
get_param = clf.get_params() # returns the hyper-parameters
predicted_array = clf.predict(
x_test
) # running the test dataset through the model, giving an array of predicted values
predic_log_proba = clf.predict_log_proba(
x_test) # log of probability estimate for each class
predic_prob = clf.predict_proba(x_test) # the probability for each class
mean_accuracy = clf.score(x_test,
y_test) # returns the mean accuracy of the test set
sparsify = clf.sparsify() # returns the coeffient matrix in sparse format
print('The mean accuracy of the test set is: %.3f' % mean_accuracy)
# now findng the confusion matrix for the data
# we first need to convert the 1 and 2 to 'female' and 'male'
