def fit(self, X, y):
# assert type(X) is pandas.DataFrame and type(y) is pandas.Series
arules_pkg = install_r_package('arulesCBA')
formula = rpy2.robjects.Formula('{} ~ .'.format(y.name))
r_train = create_r_dataframe(X, y)
hps = {k: v for k, v in self._hyperparams.items() if v is not None}
self._r_model = arules_pkg.CBA(formula=formula, data=r_train, **hps)
return self
def predict(self, X):
stats_pkg = rpy2.robjects.packages.importr('stats')
predict_fun = stats_pkg.predict
r_test = create_r_dataframe(X)
r_predictions = predict_fun(self._r_model, r_test)
levels = r_predictions.levels
predictions = [levels[yi - 1] for yi in r_predictions]
return np.array(predictions, dtype=np.int)
def fit(self, X, y):
arules_pkg = install_r_package('arulesCBA')
if not isinstance(y, pandas.Series):
y_name = "target"
else:
y_name = y.name
formula = rpy2.robjects.Formula(f'{y_name} ~ .')
r_train = create_r_dataframe(X, y)
hps = {k: v for k, v in self._hyperparams.items() if v is not None}
if False:
lale.helpers.println_pos(
'arules_pkg.CBA(formula="{}", data=[\n{}], {})'.format(
formula, r_train,
', '.join([f'{k}={v}' for k, v in hps.items()])))
self._r_model = arules_pkg.CBA(formula=formula, data=r_train, **hps)
return self