def fit(self, X, y):
n, d = X.shape
# Initial guess
self.w = np.zeros(d)
utils.check_gradient(self, X, y)
(self.w, f) = minimizers.findMin(self.funObj, self.w, self.maxEvals,
self.verbose, X, y)
def fit(self):
(self.w, self.alpha, f,
_) = minimizers.findMin(self.funObj, self.w, self.alpha,
self.maxEvals, self.verbose, self.X, self.y)
print("Training error: %.3f" %
utils.classification_error(self.predict(self.X), self.y))
def fit(self, X, y):
n, d = X.shape
self.n_classes = np.unique(y).size
self.k = np.unique(y).size
# Initial guess
self.w = np.zeros(d * self.k)
# utils.check_gradient(self, X, y)
(self.w, f) = minimizers.findMin(self.funObj, self.w, self.maxEvals,
self.verbose, X, y)
self.w = np.reshape(self.w, (d, k))
def fit(self, X, y):
n, d = X.shape
self.n_classes = np.unique(y).size
# Initial guess
self.W = np.zeros((d, self.n_classes))
for i in range(self.n_classes):
ytmp = y.copy().astype(float)
ytmp[y == i] = 1
ytmp[y != i] = -1
self.W[:, i], _ = minimizers.findMin(self.funObj, self.W[:, i],
self.maxEvals,
self.verbose,
X, ytmp)
def fit(self, X, y):
n, d = X.shape
self.n_classes = np.unique(y).size
# Initial guess
self.W = np.zeros((d, self.n_classes))
for i in range(self.n_classes):
ytmp = y.copy().astype(float)
ytmp[y == i] = 1
ytmp[y != i] = -1
# self.W[:, i] = np.linalg.lstsq(np.dot(X.T, X), np.dot(X.T, ytmp))[0]
self.w = np.zeros(d)
self.W[:,
i] = minimizers.findMin(self.funObj, self.w, self.maxEvals,
self.verbose, X, ytmp)[0]
def fit(self, X, y):
n, d = X.shape
w0 = np.zeros(d)
minimize = lambda ind: minimizers.findMin(self.funObj,
w0[ind],
self.maxEvals, 0,
X[:, ind], y)
selected = set()
selected.add(0) # always include the bias variable
minLoss = np.inf
oldLoss = 0
bestFeature = -1
while minLoss != oldLoss:
oldLoss = minLoss
if self.verbose > 1:
print("Epoch %d " % len(selected))
print("Selected feature: %d" % (bestFeature))
print("Min Loss: %.3f\n" % minLoss)
for i in range(d):
if i in selected:
continue
selected_new = selected | {i} # add "i" to the set
# TODO: Fit the model with 'i' added to the features,
# then compute the score and update the minScore/minInd
w = np.zeros(d)
w[list(selected_new)], loss = minimize(list(selected_new))
self.L0 = (self.lammy * np.count_nonzero(w))
loss += self.L0
if loss < minLoss:
minLoss = loss
bestFeature = i
w0 = w
selected.add(bestFeature)
# re-train the model one last time using the selected features
self.w = w0
self.w[list(selected)], _ = minimize(list(selected))
def fit(self, X, y):
n, d = X.shape
w0 = np.zeros(d)
minimize = lambda ind: minimizers.findMin(self.funObj, w0[ind], self.
maxEvals, 0, X[:, ind], y)
selected = set()
selected.add(0) # always include the bias variable
minLoss = np.inf
oldLoss = 0
bestFeature = -1
minScore = np.inf
minIndex = -1
# ignore = false
while minLoss != oldLoss:
oldLoss = minLoss
if self.verbose > 1:
print("Epoch %d " % len(selected))
print("Selected feature: %d" % (bestFeature))
print("Min Loss: %.3f\n" % minLoss)
for i in range(d):
if i in selected:
continue
selected_new = selected | {i} # add "i" to the set
new_w, value = minimize(list(selected_new))
# print value
if value < minScore:
minScore = value
minIndex = i
# TODO: Fit the model with 'i' added to the features,
# then compute the score and update the minScore/minInd
selected.add(minIndex)
minLoss = minScore
# re-train the model one last time using the selected features
self.w = w0
self.w[list(selected)], _ = minimize(list(selected))
def fit(self):
(self.w, self.alpha, f,
_) = minimizers.findMin(self.funObj, self.w, self.alpha,
self.maxEvals, self.verbose, self.X, self.y)
def minimize(ind):
return minimizers.findMin(self.funObj, w0[ind], self.alpha,
self.maxEvals, self.verbose,
self.X[:, ind], self.y)