def test_warm_start_path():
X, y, coef = make_user_item_regression(label_stdev=.4)
X = sp.csc_matrix(X)
rank = 4
seed = 333
step_size = 1
n_iter = 10
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)
rmse = []
for _ in range(1, n_iter):
fm.fit(X, y, n_more_iter=step_size)
rmse.append(np.sqrt(mean_squared_error(fm.predict(X), y)))
print('------- restart ----------')
rmse_re = []
for i in range(1, n_iter):
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, y)
rmse_re.append(np.sqrt(mean_squared_error(fm.predict(X), y)))
assert len(rmse) == len(rmse_re)
assert_almost_equal(rmse, rmse_re)
def test_als_warm_start():
X, y, coef = make_user_item_regression(label_stdev=0)
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_fm_regression():
w0, w, V, y, X = get_test_problem()
fm = als.FMRegression(n_iter=1000, l2_reg_w=0, l2_reg_V=0, rank=2)
fm.fit(X, y)
y_pred = fm.predict(X)
assert_almost_equal(y_pred, y, 3)
# check different size
fm = als.FMRegression(n_iter=1000, l2_reg_w=0, l2_reg_V=0, rank=5)
X_big = sp.hstack([X, X])
fm.fit(X_big, y)
y_pred = fm.predict(X_big[:2, ])
def _test_fm_regression_only_w0():
X, y = get_small_data()
fm = als.FMRegression(n_iter=0, l2_reg_w=0, l2_reg_V=0, rank=0)
fm.ignore_w = True
fm.w0_ = 2
fm.fit(X, y)
assert_almost_equal(fm.w0_, 2, 6)
fm = als.FMRegression(n_iter=1, l2_reg_w=0, l2_reg_V=0, rank=0)
fm.ignore_w = True
fm.w0_ = 2
fm.fit(X, y)
assert_almost_equal(fm.w0_, 4466.6666666666661, 6)
def test_clone():
from sklearn.base import clone
a = als.FMRegression()
b = clone(a)
assert a.get_params() == b.get_params()
a = als.FMClassification()
b = clone(a)
assert a.get_params() == b.get_params()
def test_fm_regression_reg_V():
X, y = get_small_data()
fm = als.FMRegression(n_iter=10,
l2_reg_w=1,
l2_reg_V=1,
rank=4,
random_state=123)
fm.fit(X, y)
y_pred_reg1 = fm.predict(X)
fm = als.FMRegression(n_iter=10,
l2_reg_w=1,
l2_reg_V=100,
rank=4,
random_state=123)
fm.fit(X, y)
y_pred_reg100 = fm.predict(X)
# different reg values should result in differnt predictions
assert np.any(np.not_equal(y_pred_reg1, y_pred_reg100))
def _test_raise_when_input_is_dense():
fm = als.FMRegression(n_iter=0, l2_reg_w=0, l2_reg_V=0, rank=0)
X = np.arange(3, 4, dtype=np.float64)
y = np.arange(3, dtype=np.float64)
fm.fit(X, y)
