best_lambda = 1000.
best_thresh = 0.5
best_eta = 0.000251188643151
eps = 1.0e-2
seed = 42
frac = 0.1
num = 6
lammax = 1000.0
scale = 10.0
sparse = False
# Load in MNIST training data, set 2s -> 1, others -> 0
print("Loading MNIST Training data...")
X_train, y_train = mu.load_mnist(dataset='training')
y_train_true = mu.mnist_filter(y_train, filterby=2)
print("True number of twos in training set:",np.sum(y_train_true))
# Load in MNIST training data, set 2s -> 1, others -> 0
print("Loading MNIST Testing data...")
X_test, y_test = mu.load_mnist(dataset='testing')
y_test_true = mu.mnist_filter(y_test, filterby=2)
# Build regularized binary classifier by minimizing 01 loss
# Will need to optimize over lambda and eta via a regularization path
if find_best_lam:
print("Running regularization path to optmize lambda, eta...")
# Split training data into subtraining set, validation set
X_tr, y_tr, X_val, y_val = val.split_data(X_train, y_train, frac=frac, seed=seed)
kwargs = {}
# Fit for 4s
# Perform grid search to find best regularization constant and threshold?
if find_best_lam:
print("Finding optimal lambda and threshold via regularization path.")
thresh_arr = np.linspace(-0.2,1.,num)
err_val = np.zeros((num,num))
err_train = np.zeros((num,num))
# Split training data into subtraining set, validation set
X_tr, y_tr, X_val, y_val = val.split_data(X_train, y_train, frac=frac, seed=seed)
# Filter y values to 0, 1 labels
y_tr_true = mu.mnist_filter(y_tr,filterby=3)
y_val_true = mu.mnist_filter(y_val,filterby=3)
# Loop over thresholds
for i in range(num):
# Internally loop over lambdas in regularization path
err_val[i,:], err_train[i,:], lams = val.linear_reg_path(X_tr, y_tr_true, X_val,
y_val_true, ri.fit_ridge,
lammax=lammax,
scale=scale,
num=num, error_func=val.loss_01,
thresh=thresh_arr[i], **kwargs)
# Find minimum threshold, lambda from minimum validation error
ind_t,ind_l = np.unravel_index(err_val.argmin(), err_val.shape)
best_lambda = lams[ind_l]
# Perform grid search to find best regularization constant and threshold?
if find_best_lam:
print("Finding optimal lambda and threshold via regularization path.")
thresh_arr = np.linspace(-0.2, 1., num)
err_val = np.zeros((num, num))
err_train = np.zeros((num, num))
# Split training data into subtraining set, validation set
X_tr, y_tr, X_val, y_val = val.split_data(X_train,
y_train,
frac=frac,
seed=seed)
# Filter y values to 0, 1 labels
y_tr_true = mu.mnist_filter(y_tr, filterby=3)
y_val_true = mu.mnist_filter(y_val, filterby=3)
# Loop over thresholds
for i in range(num):
# Internally loop over lambdas in regularization path
err_val[i, :], err_train[i, :], lams = val.linear_reg_path(
X_tr,
y_tr_true,
X_val,
y_val_true,
ri.fit_ridge,
lammax=lammax,
scale=scale,
num=num,
error_func=val.loss_01,
best_lambda = 1000.
best_thresh = 0.5
best_eta = 0.000251188643151
eps = 1.0e-2
seed = 42
frac = 0.1
num = 6
lammax = 1000.0
scale = 10.0
sparse = False
# Load in MNIST training data, set 2s -> 1, others -> 0
print("Loading MNIST Training data...")
X_train, y_train = mu.load_mnist(dataset='training')
y_train_true = mu.mnist_filter(y_train, filterby=2)
print("True number of twos in training set:", np.sum(y_train_true))
# Load in MNIST training data, set 2s -> 1, others -> 0
print("Loading MNIST Testing data...")
X_test, y_test = mu.load_mnist(dataset='testing')
y_test_true = mu.mnist_filter(y_test, filterby=2)
# Build regularized binary classifier by minimizing 01 loss
# Will need to optimize over lambda and eta via a regularization path
if find_best_lam:
print("Running regularization path to optmize lambda, eta...")
# Split training data into subtraining set, validation set
X_tr, y_tr, X_val, y_val = val.split_data(X_train,
y_train,
