def test_make_user_item_regression():
from fastFM.mcmc import FMRegression
X, y, coef = make_user_item_regression(label_stdev=0)
from sklearn.cross_validation import train_test_split
X_train, X_test, y_train, y_test = train_test_split(X,
y,
test_size=0.33,
random_state=42)
fm = FMRegression(rank=2)
y_pred = fm.fit_predict(sp.csc_matrix(X_train), y_train,
sp.csc_matrix(X_test))
# generate data with noisy lables
X, y, coef = make_user_item_regression(label_stdev=2)
from sklearn.cross_validation import train_test_split
X_train, X_test, y_train, y_test = train_test_split(X,
y,
test_size=0.33,
random_state=42)
fm = FMRegression(rank=2)
y_pred_noise = fm.fit_predict(sp.csc_matrix(X_train), y_train,
sp.csc_matrix(X_test))
assert mean_squared_error(y_pred_noise, y_test) > \
mean_squared_error(y_pred, y_test)
def test_find_init_stdev():
X, y, coef = make_user_item_regression(label_stdev=.5)
from sklearn.model_selection import train_test_split
X_train, X_test, y_train, y_test = train_test_split(X,
y,
test_size=0.33,
random_state=44)
X_train = sp.csc_matrix(X_train)
X_test = sp.csc_matrix(X_test)
fm = mcmc.FMRegression(n_iter=10, rank=5)
best_init_stdev, mse = mcmc.find_init_stdev(fm,
X_train,
y_train,
stdev_range=[0.2, 0.5, 1.0])
best_init_stdev_bad, _ = mcmc.find_init_stdev(fm,
X_train,
y_train,
stdev_range=[5.])
print('--' * 30)
best_init_stdev_vali, mse_vali = mcmc.find_init_stdev(
fm, X_train, y_train, X_test, y_test, stdev_range=[0.2, 0.5, 1.0])
assert best_init_stdev < best_init_stdev_bad
assert best_init_stdev_vali == best_init_stdev
assert mse_vali > mse
def test_als_warm_start():
X, y, coef = make_user_item_regression(label_stdev=0)
from sklearn.cross_validation import train_test_split
X_train, X_test, y_train, y_test = train_test_split(
X, y, test_size=0.33, random_state=42)
X_train = sp.csc_matrix(X_train)
X_test = sp.csc_matrix(X_test)
fm = als.FMRegression(n_iter=10, l2_reg_w=0, l2_reg_V=0, rank=2)
fm.fit(X_train, y_train)
y_pred = fm.predict(X_test)
error_10_iter = mean_squared_error(y_pred, y_test)
fm = als.FMRegression(n_iter=5, l2_reg_w=0, l2_reg_V=0, rank=2)
fm.fit(X_train, y_train)
print(fm.iter_count)
y_pred = fm.predict(X_test)
error_5_iter = mean_squared_error(y_pred, y_test)
fm.fit(sp.csc_matrix(X_train), y_train, n_more_iter=5)
print(fm.iter_count)
y_pred = fm.predict(X_test)
error_5_iter_plus_5 = mean_squared_error(y_pred, y_test)
print(error_5_iter, error_5_iter_plus_5, error_10_iter)
assert error_10_iter == error_5_iter_plus_5
def test_als_warm_start():
X, y, coef = make_user_item_regression(label_stdev=0)
from sklearn.cross_validation import train_test_split
X_train, X_test, y_train, y_test = train_test_split(
X, y, test_size=0.33, random_state=42)
X_train = sp.csc_matrix(X_train)
X_test = sp.csc_matrix(X_test)
fm = als.FMRegression(n_iter=10, l2_reg_w=0, l2_reg_V=0, rank=2)
fm.fit(X_train, y_train)
y_pred = fm.predict(X_test)
error_10_iter = mean_squared_error(y_pred, y_test)
fm = als.FMRegression(n_iter=5, l2_reg_w=0, l2_reg_V=0, rank=2)
fm.fit(X_train, y_train)
print fm.iter_count
y_pred = fm.predict(X_test)
error_5_iter = mean_squared_error(y_pred, y_test)
fm.fit(sp.csc_matrix(X_train), y_train, n_more_iter=5)
print fm.iter_count
y_pred = fm.predict(X_test)
error_5_iter_plus_5 = mean_squared_error(y_pred, y_test)
print error_5_iter, error_5_iter_plus_5, error_10_iter
assert error_10_iter == error_5_iter_plus_5
def test_make_user_item_regression():
from fastFM.mcmc import FMRegression
X, y, coef = make_user_item_regression(label_stdev=0)
from sklearn.cross_validation import train_test_split
X_train, X_test, y_train, y_test = train_test_split(
X, y, test_size=0.33, random_state=42)
fm = FMRegression(rank=2)
y_pred = fm.fit_predict(sp.csc_matrix(X_train), y_train, sp.csc_matrix(X_test))
# generate data with noisy lables
X, y, coef = make_user_item_regression(label_stdev=2)
from sklearn.cross_validation import train_test_split
X_train, X_test, y_train, y_test = train_test_split(
X, y, test_size=0.33, random_state=42)
fm = FMRegression(rank=2)
y_pred_noise = fm.fit_predict(sp.csc_matrix(X_train), y_train, sp.csc_matrix(X_test))
assert mean_squared_error(y_pred_noise, y_test) > \
mean_squared_error(y_pred, y_test)
def test_warm_start_path():
X, y, coef = make_user_item_regression(label_stdev=.4)
from sklearn.cross_validation import train_test_split
X_train, X_test, y_train, y_test = train_test_split(X,
y,
test_size=0.33,
random_state=42)
X_train = sp.csc_matrix(X_train)
X_test = sp.csc_matrix(X_test)
n_iter = 10
rank = 4
seed = 333
step_size = 1
l2_reg_w = 0
l2_reg_V = 0
fm = als.FMRegression(n_iter=0,
l2_reg_w=l2_reg_w,
l2_reg_V=l2_reg_V,
rank=rank,
random_state=seed)
# initalize coefs
fm.fit(X_train, y_train)
rmse_train = []
rmse_test = []
for i in range(1, n_iter):
fm.fit(X_train, y_train, n_more_iter=step_size)
rmse_train.append(
np.sqrt(mean_squared_error(fm.predict(X_train), y_train)))
rmse_test.append(
np.sqrt(mean_squared_error(fm.predict(X_test), y_test)))
print('------- restart ----------')
values = np.arange(1, n_iter)
rmse_test_re = []
rmse_train_re = []
for i in values:
fm = als.FMRegression(n_iter=i,
l2_reg_w=l2_reg_w,
l2_reg_V=l2_reg_V,
rank=rank,
random_state=seed)
fm.fit(X_train, y_train)
rmse_test_re.append(
np.sqrt(mean_squared_error(fm.predict(X_test), y_test)))
rmse_train_re.append(
np.sqrt(mean_squared_error(fm.predict(X_train), y_train)))
assert_almost_equal(rmse_train, rmse_train_re)
assert_almost_equal(rmse_test, rmse_test_re)
def test_warm_start_path():
X, y, coef = make_user_item_regression(label_stdev=.4)
from sklearn.model_selection import train_test_split
X_train, X_test, y_train, y_test = train_test_split(
X, y, test_size=0.33, random_state=42)
X_train = sp.csc_matrix(X_train)
X_test = sp.csc_matrix(X_test)
n_iter = 10
rank = 4
seed = 333
step_size = 1
l2_reg_w = 0
l2_reg_V = 0
fm = als.FMRegression(n_iter=0, l2_reg_w=l2_reg_w,
l2_reg_V=l2_reg_V, rank=rank, random_state=seed)
# initalize coefs
fm.fit(X_train, y_train)
rmse_train = []
rmse_test = []
for i in range(1, n_iter):
fm.fit(X_train, y_train, n_more_iter=step_size)
rmse_train.append(np.sqrt(mean_squared_error(
fm.predict(X_train), y_train)))
rmse_test.append(np.sqrt(mean_squared_error(
fm.predict(X_test), y_test)))
print('------- restart ----------')
values = np.arange(1, n_iter)
rmse_test_re = []
rmse_train_re = []
for i in values:
fm = als.FMRegression(n_iter=i, l2_reg_w=l2_reg_w,
l2_reg_V=l2_reg_V, rank=rank, random_state=seed)
fm.fit(X_train, y_train)
rmse_test_re.append(np.sqrt(mean_squared_error(
fm.predict(X_test), y_test)))
rmse_train_re.append(np.sqrt(mean_squared_error(
fm.predict(X_train), y_train)))
assert_almost_equal(rmse_train, rmse_train_re)
assert_almost_equal(rmse_test, rmse_test_re)
def test_find_init_stdev():
X, y, coef = make_user_item_regression(label_stdev=.5)
from sklearn.cross_validation import train_test_split
X_train, X_test, y_train, y_test = train_test_split(
X, y, test_size=0.33, random_state=44)
X_train = sp.csc_matrix(X_train)
X_test = sp.csc_matrix(X_test)
fm = mcmc.FMRegression(n_iter=10, rank=5)
best_init_stdev, mse = mcmc.find_init_stdev(fm, X_train, y_train,
stdev_range=[0.2, 0.5, 1.0])
best_init_stdev_bad, _ = mcmc.find_init_stdev(fm, X_train, y_train,
stdev_range=[5.])
print('--' * 30)
best_init_stdev_vali, mse_vali = mcmc.find_init_stdev(fm, X_train, y_train, X_test,
y_test, stdev_range=[0.2, 0.5, 1.0])
assert best_init_stdev < best_init_stdev_bad
assert best_init_stdev_vali == best_init_stdev
assert mse_vali > mse
def test_mcmc_warm_start():
X, y, coef = make_user_item_regression(label_stdev=0)
from sklearn.model_selection import train_test_split
X_train, X_test, y_train, y_test = train_test_split(
X, y, test_size=0.33, random_state=44)
X_train = sp.csc_matrix(X_train)
X_test = sp.csc_matrix(X_test)
fm = mcmc.FMRegression(n_iter=100, rank=2)
y_pred = fm.fit_predict(X_train, y_train, X_test)
error_10_iter = mean_squared_error(y_pred, y_test)
fm = mcmc.FMRegression(n_iter=50, rank=2)
y_pred = fm.fit_predict(X_train, y_train, X_test)
error_5_iter = mean_squared_error(y_pred, y_test)
y_pred = fm.fit_predict(X_train, y_train, X_test, n_more_iter=50)
error_5_iter_plus_5 = mean_squared_error(y_pred, y_test)
print(error_5_iter, error_5_iter_plus_5, error_10_iter)
print(fm.hyper_param_)
assert_almost_equal(error_10_iter, error_5_iter_plus_5, decimal=2)
def test_mcmc_warm_start():
X, y, coef = make_user_item_regression(label_stdev=0)
from sklearn.cross_validation import train_test_split
X_train, X_test, y_train, y_test = train_test_split(
X, y, test_size=0.33, random_state=44)
X_train = sp.csc_matrix(X_train)
X_test = sp.csc_matrix(X_test)
fm = mcmc.FMRegression(n_iter=100, rank=2)
y_pred = fm.fit_predict(X_train, y_train, X_test)
error_10_iter = mean_squared_error(y_pred, y_test)
fm = mcmc.FMRegression(n_iter=50, rank=2)
y_pred = fm.fit_predict(X_train, y_train, X_test)
error_5_iter = mean_squared_error(y_pred, y_test)
y_pred = fm.fit_predict(X_train, y_train, X_test, n_more_iter=50)
error_5_iter_plus_5 = mean_squared_error(y_pred, y_test)
print(error_5_iter, error_5_iter_plus_5, error_10_iter)
print(fm.hyper_param_)
assert_almost_equal(error_10_iter, error_5_iter_plus_5, decimal=2)
from fastFM.datasets import make_user_item_regression
from sklearn.cross_validation import train_test_split
# This sets up a small test dataset.
X, y, _ = make_user_item_regression(label_stdev=.4)
X_train, X_test, y_train, y_test = train_test_split(X, y)
from fastFM import als
fm = als.FMRegression(n_iter=1000, init_stdev=0.1, rank=2, l2_reg_w=0.1, l2_reg_V=0.5)
fm.fit(X_train, y_train)
y_pred = fm.predict(X_test)
from sklearn.metrics import mean_squared_error
print('mse:', mean_squared_error(y_test, y_pred))
import numpy as np
# Convert dataset to binary classification task.
y_labels = np.ones_like(y)
y_labels[y < np.mean(y)] = -1
X_train, X_test, y_train, y_test = train_test_split(X, y_labels)
from fastFM import sgd
fm = sgd.FMClassification(n_iter=1000, init_stdev=0.1, l2_reg_w=0,
l2_reg_V=0, rank=2, step_size=0.1)
fm.fit(X_train, y_train)
y_pred = fm.predict(X_test)
y_pred_proba = fm.predict_proba(X_test)
from sklearn.metrics import accuracy_score, roc_auc_score
print('acc:', accuracy_score(y_test, y_pred))
