def train(self, shared=True):
ucifolder = UCIFolder(self.D, normalize=False, shuffle=False)
self.X_train, self.Y_train = [], []
self.X_test, self.Y_test, self.P = [], [], []
for c in [80]:#[5,10,15,20,30,50,80]:
# Get data and labels at fold k
X,Y = ucifolder.training(c)
print X.shape, Y.shape
# Get the testing data
Xi,Yi = ucifolder.testing(c)
# Solve for the vector of linear factors, W
return self.boost(X, Y, Xi, Yi, self.thresh(X))
def train(self, shared=True):
ucifolder = UCIFolder(self.D, normalize=False, shuffle=True)
self.X_train, self.Y_train = [], []
self.X_test, self.Y_test, self.P = [], [], []
for c in [5,10,15,20,30,50,80]:
# Get data and labels at fold k
X,Y = ucifolder.training(c)
# Get the testing data
Xi,Yi = ucifolder.testing(c)
# Solve for the vector of linear factors, W
train_error, test_error, test_auc = self.boost(X, Y, Xi, Yi, self.thresh(X))
print "c%="+str(c)+"%, train error:", "%.2f" % train_error,
print "test error:", "%.2f" % test_error, "AUC:", "%.2f" % test_auc
def train(self, shared=True):
ucifolder = UCIFolder(self.D, normalize=False, shuffle=False)
n = len(self.D) / 50
t = len(self.D) / 2
for c in [5]: # [5,10,15,20,30,50]:
# Get the active learning pool
px, py = ucifolder.training(c, pool=True)
# Get data and labels at fold k
X, Y = ucifolder.training(c)
# Get the testing data
Xi, Yi = ucifolder.testing(c)
# Do active learning
while len(X) < t:
print "Training pool size:", float(len(X)) / float(len(self.D))
classifier = self.boost(X, Y, Xi, Yi, self.thresh(X))
X, Y, px, py = self.add(X, Y, px, py, classifier, n)
