def predict(self,
X,
p,
treatment=None,
y=None,
return_components=False,
verbose=True):
"""Predict treatment effects.
Args:
X (np.matrix): a feature matrix
p (np.ndarray or dict): an array of propensity scores of float (0,1) in the single-treatment case
or, a dictionary of treatment groups that map to propensity vectors of float (0,1)
treatment (np.array, optional): a treatment vector
y (np.array, optional): an optional outcome vector
return_componets (bool, optional): whether to return outcome for treatment and control seperately
Returns:
(numpy.ndarray): Predictions of treatment effects.
"""
check_p_conditions(p, self.t_groups)
if isinstance(p, np.ndarray):
treatment_name = self.t_groups[0]
p = {treatment_name: p}
te = np.zeros((X.shape[0], self.t_groups.shape[0]))
dhat_cs = {}
dhat_ts = {}
for i, group in enumerate(self.t_groups):
model_tau_c = self.models_tau_c[group]
model_tau_t = self.models_tau_t[group]
dhat_cs[group] = model_tau_c.predict(X)
dhat_ts[group] = model_tau_t.predict(X)
_te = (p[group] * dhat_cs[group] +
(1 - p[group]) * dhat_ts[group]).reshape(-1, 1)
te[:, i] = np.ravel(_te)
if (y is not None) and (treatment is not None) and verbose:
mask = (treatment == group) | (treatment == self.control_name)
treatment_filt = treatment[mask]
X_filt = X[mask]
y_filt = y[mask]
w = (treatment_filt == group).astype(int)
yhat = np.zeros_like(y_filt, dtype=float)
yhat[w == 0] = self.models_mu_c[group].predict(X_filt[w == 0])
yhat[w == 1] = self.models_mu_t[group].predict(X_filt[w == 1])
logger.info('Error metrics for group {}'.format(group))
regression_metrics(y_filt, yhat, w)
if not return_components:
return te
else:
return te, dhat_cs, dhat_ts
def predict(self,
X,
treatment=None,
y=None,
return_components=False,
verbose=True):
"""Predict treatment effects.
Args:
X (np.matrix or np.array or pd.Dataframe): a feature matrix
treatment (np.array or pd.Series, optional): a treatment vector
y (np.array or pd.Series, optional): an outcome vector
return_components (bool, optional): whether to return outcome for treatment and control seperately
verbose (bool, optional): whether to output progress logs
Returns:
(numpy.ndarray): Predictions of treatment effects.
"""
X, treatment, y = convert_pd_to_np(X, treatment, y)
yhat_cs = {}
yhat_ts = {}
for group in self.t_groups:
model = self.models[group]
# set the treatment column to zero (the control group)
X_new = np.hstack((np.zeros((X.shape[0], 1)), X))
yhat_cs[group] = model.predict(X_new)
# set the treatment column to one (the treatment group)
X_new[:, 0] = 1
yhat_ts[group] = model.predict(X_new)
if (y is not None) and (treatment is not None) and verbose:
mask = (treatment == group) | (treatment == self.control_name)
treatment_filt = treatment[mask]
w = (treatment_filt == group).astype(int)
y_filt = y[mask]
yhat = np.zeros_like(y_filt, dtype=float)
yhat[w == 0] = yhat_cs[group][mask][w == 0]
yhat[w == 1] = yhat_ts[group][mask][w == 1]
logger.info('Error metrics for group {}'.format(group))
regression_metrics(y_filt, yhat, w)
te = np.zeros((X.shape[0], self.t_groups.shape[0]))
for i, group in enumerate(self.t_groups):
te[:, i] = yhat_ts[group] - yhat_cs[group]
if not return_components:
return te
else:
return te, yhat_cs, yhat_ts
return te
def predict(self,
X,
treatment=None,
y=None,
p=None,
return_components=False,
verbose=True):
"""Predict treatment effects.
Args:
X (np.matrix or np.array or pd.Dataframe): a feature matrix
treatment (np.array or pd.Series, optional): a treatment vector
y (np.array or pd.Series, optional): an outcome vector
verbose (bool, optional): whether to output progress logs
Returns:
(numpy.ndarray): Predictions of treatment effects.
"""
X, treatment, y = convert_pd_to_np(X, treatment, y)
te = np.zeros((X.shape[0], self.t_groups.shape[0]))
yhat_cs = {}
yhat_ts = {}
for i, group in enumerate(self.t_groups):
models_tau = self.models_tau[group]
_te = np.r_[[model.predict(X)
for model in models_tau]].mean(axis=0)
te[:, i] = np.ravel(_te)
yhat_cs[group] = np.r_[[
model.predict(X) for model in self.models_mu_c
]].mean(axis=0)
yhat_ts[group] = np.r_[[
model.predict(X) for model in self.models_mu_t[group]
]].mean(axis=0)
if (y is not None) and (treatment is not None) and verbose:
mask = (treatment == group) | (treatment == self.control_name)
treatment_filt = treatment[mask]
X_filt = X[mask]
y_filt = y[mask]
w = (treatment_filt == group).astype(int)
yhat = np.zeros_like(y_filt, dtype=float)
yhat[w == 0] = yhat_cs[group][mask][w == 0]
yhat[w == 1] = yhat_ts[group][mask][w == 1]
logger.info("Error metrics for group {}".format(group))
regression_metrics(y_filt, yhat, w)
if not return_components:
return te
else:
return te, yhat_cs, yhat_ts
def predict(self,
X,
treatment=None,
y=None,
return_components=False,
verbose=True):
"""Predict treatment effects.
Args:
X (np.matrix): a feature matrix
treatment (np.array, optional): a treatment vector
y (np.array, optional): an optional outcome vector
return_componets (bool, optional): whether to return outcome for treatment and control seperately
Returns:
(numpy.ndarray): Predictions of treatment effects.
"""
yhat_cs = {}
yhat_ts = {}
for group in self.t_groups:
model_c = self.models_c[group]
model_t = self.models_t[group]
yhat_cs[group] = model_c.predict(X)
yhat_ts[group] = model_t.predict(X)
if (y is not None) and (treatment is not None) and verbose:
mask = (treatment == group) | (treatment == self.control_name)
treatment_filt = treatment[mask]
y_filt = y[mask]
w = (treatment_filt == group).astype(int)
yhat = np.zeros_like(y_filt, dtype=float)
yhat[w == 0] = yhat_cs[group][mask][w == 0]
yhat[w == 1] = yhat_ts[group][mask][w == 1]
logger.info('Error metrics for group {}'.format(group))
regression_metrics(y_filt, yhat, w)
te = np.zeros((X.shape[0], self.t_groups.shape[0]))
for i, group in enumerate(self.t_groups):
te[:, i] = yhat_ts[group] - yhat_cs[group]
if not return_components:
return te
else:
return te, yhat_cs, yhat_ts
def predict(self, X, treatment=None, y=None, verbose=True):
"""Predict treatment effects.
Args:
X (np.matrix): a feature matrix
treatment (np.array, optional): a treatment vector
y (np.array, optional): an outcome vector
Returns:
(numpy.ndarray): Predictions of treatment effects.
"""
yhat_cs = {}
yhat_ts = {}
for group in self.t_groups:
model = self.models[group]
# set the treatment column to zero (the control group)
X_new = np.hstack((np.zeros((X.shape[0], 1)), X))
yhat_cs[group] = model.predict(X_new)
# set the treatment column to one (the treatment group)
X_new[:, 0] = 1
yhat_ts[group] = model.predict(X_new)
if (y is not None) and (treatment is not None) and verbose:
mask = (treatment == group) | (treatment == self.control_name)
treatment_filt = treatment[mask]
w = (treatment_filt == group).astype(int)
y_filt = y[mask]
yhat = np.zeros_like(y_filt, dtype=float)
yhat[w == 0] = yhat_cs[group][mask][w == 0]
yhat[w == 1] = yhat_ts[group][mask][w == 1]
logger.info('Error metrics for group {}'.format(group))
regression_metrics(y_filt, yhat, w)
te = np.zeros((X.shape[0], self.t_groups.shape[0]))
for i, group in enumerate(self.t_groups):
te[:, i] = yhat_ts[group] - yhat_cs[group]
return te
def predict(self, X, treatment=None, y=None, return_components=False, verbose=True):
"""Predict treatment effects.
Args:
X (np.matrix): a feature matrix
treatment (np.array, optional): a treatment vector
y (np.array, optional): an optional outcome vector
Returns:
(numpy.ndarray): Predictions of treatment effects.
"""
yhat_cs = {}
yhat_ts = {}
for group in self.t_groups:
model_c = self.models_c[group]
model_t = self.models_t[group]
yhat_cs[group] = model_c.predict(X)
yhat_ts[group] = model_t.predict(X)
if (y is not None) and (treatment is not None) and verbose:
w = (treatment == group)
yhat = np.zeros_like(y, dtype=float)
yhat[~w] = yhat_cs[group][~w]
yhat[w] = yhat_ts[group][w]
logger.info('Error metrics for group {}'.format(group))
regression_metrics(y, yhat, w)
te = np.zeros((X.shape[0], self.t_groups.shape[0]))
for i, group in enumerate(self.t_groups):
te[:, i] = yhat_ts[group] - yhat_cs[group]
if not return_components:
return te
else:
return te, yhat_cs, yhat_ts
def predict(self,
X,
treatment=None,
y=None,
p=None,
return_components=False,
verbose=True):
"""Predict treatment effects.
Args:
X (np.matrix or np.array or pd.Dataframe): a feature matrix
treatment (np.array or pd.Series, optional): a treatment vector
y (np.array or pd.Series, optional): an outcome vector
p (np.ndarray or pd.Series or dict, optional): an array of propensity scores of float (0,1) in the
single-treatment case; or, a dictionary of treatment groups that map to propensity vectors of
float (0,1); if None will run ElasticNetPropensityModel() to generate the propensity scores.
return_components (bool, optional): whether to return outcome for treatment and control seperately
verbose (bool, optional): whether to output progress logs
Returns:
(numpy.ndarray): Predictions of treatment effects.
"""
X, treatment, y = convert_pd_to_np(X, treatment, y)
if p is None:
logger.info('Generating propensity score')
p = dict()
for group in self.t_groups:
p_model = self.propensity_model[group]
p[group] = p_model.predict(X)
else:
check_p_conditions(p, self.t_groups)
if isinstance(p, (np.ndarray, pd.Series)):
treatment_name = self.t_groups[0]
p = {treatment_name: convert_pd_to_np(p)}
elif isinstance(p, dict):
p = {
treatment_name: convert_pd_to_np(_p)
for treatment_name, _p in p.items()
}
te = np.zeros((X.shape[0], self.t_groups.shape[0]))
dhat_cs = {}
dhat_ts = {}
for i, group in enumerate(self.t_groups):
model_tau_c = self.models_tau_c[group]
model_tau_t = self.models_tau_t[group]
dhat_cs[group] = model_tau_c.predict(X)
dhat_ts[group] = model_tau_t.predict(X)
_te = (p[group] * dhat_cs[group] +
(1 - p[group]) * dhat_ts[group]).reshape(-1, 1)
te[:, i] = np.ravel(_te)
if (y is not None) and (treatment is not None) and verbose:
mask = (treatment == group) | (treatment == self.control_name)
treatment_filt = treatment[mask]
X_filt = X[mask]
y_filt = y[mask]
w = (treatment_filt == group).astype(int)
yhat = np.zeros_like(y_filt, dtype=float)
yhat[w == 0] = self.models_mu_c[group].predict(X_filt[w == 0])
yhat[w == 1] = self.models_mu_t[group].predict(X_filt[w == 1])
logger.info('Error metrics for group {}'.format(group))
regression_metrics(y_filt, yhat, w)
if not return_components:
return te
else:
return te, dhat_cs, dhat_ts
