def test_model_specification_error(self):
n_l, prior_l = 30, .3
n_u, prior_u = 200, .5
n_t = 100
x, y, x_tp, x_tn = gendata(n_l, prior_l, n_u, prior_u, n_t)
eta_list = [
pnu_mr.calc_etab(np.sum(y == +1), np.sum(y == -1), prior_u)
]
with self.assertRaises(ValueError):
pnu_mr.PNU_SL_FastCV(x, y, prior_u, eta_list, basis='DNN')
def test_calc_etab(self):
n_p, n_n = 10, 10
prior_list = np.arange(.1, 1, .1)
for prior in prior_list:
eta = pnu_mr.calc_etab(n_p, n_n, prior)
if prior < .5:
self.assertTrue(eta > 0)
elif prior > .5:
self.assertTrue(eta < 0)
else:
self.assertTrue(eta == 0)
def test_label_specification_error(self):
n_l, prior_l = 30, .3
n_u, prior_u = 200, .5
n_t = 100
x, y, x_tp, x_tn = gendata(n_l, prior_l, n_u, prior_u, n_t)
y[0] = 2
eta_list = [
pnu_mr.calc_etab(np.sum(y == +1), np.sum(y == -1), prior_u)
]
with self.assertRaises(ValueError):
pnu_mr.PNU_SL(x, y, prior_u, eta_list, model='lm')
def test_linear_model_one_eta(self):
n_l, prior_l = 30, .5
n_u, prior_u = 200, .3
n_t = 100
x, y, x_tp, x_tn = gendata(n_l, prior_l, n_u, prior_u, n_t)
eta_list = [
pnu_mr.calc_etab(np.sum(y == +1), np.sum(y == -1), prior_u)
]
f_dec, outs, funcs = pnu_mr.PNU_SL_FastCV(x,
y,
prior_u,
eta_list,
basis='lm',
nargout=3)
err = 100 * calc_err(f_dec, x_tp, x_tn, prior_u)
def test_model_nargout2(self):
n_l, prior_l = 30, .5
n_u, prior_u = 200, .3
n_t = 100
x, y, x_tp, x_tn = gendata(n_l, prior_l, n_u, prior_u, n_t)
eta_list = [
pnu_mr.calc_etab(np.sum(y == +1), np.sum(y == -1), prior_u)
]
lambda_list = [.1]
f_dec, outs = pnu_mr.PNU_SL_FastCV(x,
y,
prior_u,
eta_list,
lambda_list=lambda_list,
basis='lm',
nargout=2)
err = 100 * calc_err(f_dec, x_tp, x_tn, prior_u)
def test_gauss_model(self):
n_l, prior_l = 30, .3
n_u, prior_u = 200, .5
n_t = 100
lambda_list = [.1]
x, y, x_tp, x_tn = gendata(n_l, prior_l, n_u, prior_u, n_t)
eta_list = [
pnu_mr.calc_etab(np.sum(y == +1), np.sum(y == -1), prior_u)
]
f_dec, outs, funcs = pnu_mr.PNU_SL_FastCV(x,
y,
prior_u,
eta_list,
basis='gauss',
lambda_list=lambda_list,
nargout=3)
err = 100 * calc_err(f_dec, x_tp, x_tn, prior_u)
def test_model_without_cv(self):
n_l, prior_l = 30, .3
n_u, prior_u = 200, .5
n_t = 100
x, y, x_tp, x_tn = gendata(n_l, prior_l, n_u, prior_u, n_t)
eta_list = [
pnu_mr.calc_etab(np.sum(y == +1), np.sum(y == -1), prior_u)
]
lambda_list = [.1]
f_dec, outs, funcs = pnu_mr.PNU_SL(x,
y,
prior_u,
eta_list,
lambda_list=lambda_list,
model='lm',
nargout=3)
err = 100 * calc_err(f_dec, x_tp, x_tn, prior_u)
n_l = 20
n_u = 1000
n_t = 1000
prior_l = .5
prior_u = .3
eta_list = np.mgrid[-.9:.9:19j]
x, y, x_t, y_t = gen_twonorm_ssl(n_l, prior_l, n_u, prior_u, n_t)
x_l, y_l, x_u = x[y != 0, :], y[y != 0], x[y == 0, :]
priorh = cpe(x_l, y_l, x_u)
clf = PNU_SL(prior=priorh, basis='lm')
params = {'eta': eta_list, 'lam': [.1]}
etah = calc_etab(np.sum(y == +1), np.sum(y == -1), priorh)
clf = GridSearchCV(estimator=clf,
param_grid=params,
scoring=make_scorer(pnu_risk, prior=priorh, eta=etah),
cv=3,
n_jobs=-1)
timer.tic("Start PNU_SL")
clf.fit(x, y)
timer.toc()
y_h = clf.predict(x_t)
print(y_h)
print(y_t)
err1 = 100 * bin_clf_err(y_h, y_t, prior=prior_u)
print("Error: {:.2f}\n".format(err1))
timer.tic("Start PNU_SL_FastCV")