def test_model(model,
X,
y,
error_fn=None,
score_fn=None,
search_params=None,
verbose=True):
maybe_print('Training model...', verbose)
if search_params is not None:
validator = cv.DataFrameCV(model,
search_params,
n_folds=args.nfolds,
error_fn=error_fn,
score_fn=score_fn,
verbose=verbose)
validator.fit(X, y)
err = validator.best_result
else:
err = cv.cv_dataframe(model,
X,
y,
err_fn,
score_fn,
n_folds=args.nfolds,
verbose=verbose)
return err
def train_model(model, X, y, parameters=None, verbose=True):
if parameters is None:
parameters = {}
else:
parameters = parse_parameters(parameters)
model.set_params(**parameters)
maybe_print('Fitting model...', verbose)
model.fit(X, y)
maybe_print('Done!', verbose)
return model
def train_model(model, X, y, parameters=None, verbose=True):
if parameters is None:
parameters = {}
else:
parameters = parse_parameters(parameters)
model.set_params(**parameters)
maybe_print('Fitting model...', verbose)
model.fit(X, y)
maybe_print('Done!', verbose)
return model
def test_model(model, X, y, error_fn=None, score_fn=None,
search_params=None, verbose=True):
maybe_print('Training model...', verbose)
if search_params is not None:
validator = cv.DataFrameCV(model, search_params,
n_folds=args.nfolds,
error_fn=error_fn,
score_fn=score_fn,
verbose=verbose)
validator.fit(X, y)
err = validator.best_result
else:
err = cv.cv_dataframe(model, X, y, err_fn, score_fn,
n_folds=args.nfolds, verbose=verbose)
return err
def setup_model(source, builder, objective_fn, verbose=True):
maybe_print('Building events...', verbose)
X = builder.build_events(source)
maybe_print('Calculating objective...', verbose)
y = objective_fn(X)
return (X, y)
def setup_model(source, builder, objective_fn, verbose=True):
maybe_print('Building events...', verbose)
X = builder.build_events(source)
maybe_print('Calculating objective...', verbose)
y = objective_fn(X)
return (X, y)