def _get_model_cv_preds(self, model, X_train, y_train, cache_file):
"""
Return cross-validation predictions on the training set, using cache
if possible.
This is used if stacking is enabled (ie. a second model is used to
combine the stage 0 predictions).
"""
stack_preds = load_from_cache(
"cache/models/%s/cv_preds/%s.pkl" % (self.cache_dir, cache_file),
self.use_cached_models)
if stack_preds is None:
kfold = cross_validation.StratifiedKFold(y_train, 4)
stack_preds = []
indexes_cv = []
for stage0, stack in kfold:
model.fit(X_train[stage0], y_train[stage0])
stack_preds.extend(list(model.predict_proba(
X_train[stack])[:, 1]))
indexes_cv.extend(list(stack))
stack_preds = np.array(stack_preds)[sp.argsort(indexes_cv)]
with open("cache/models/%s/cv_preds/%s.pkl" % (
self.cache_dir, cache_file), 'w+') as f:
pickle.dump(stack_preds, f, pickle.HIGHEST_PROTOCOL)
return stack_preds
def _get_model_cv_preds(self, model, X_train, y_train, cache_file):
"""
Return cross-validation predictions on the training set, using cache
if possible.
This is used if stacking is enabled (ie. a second model is used to
combine the stage 0 predictions).
"""
stack_preds = load_from_cache(
"models/%s/cv_preds/%s.pkl" % (self.cache_dir, cache_file),
self.use_cached_models)
if stack_preds is None:
kfold = cross_validation.StratifiedKFold(y_train, 4)
stack_preds = []
indexes_cv = []
for stage0, stack in kfold:
model.fit(X_train[stage0], y_train[stage0])
stack_preds.extend(
list(model.predict_proba(X_train[stack])[:, 1]))
indexes_cv.extend(list(stack))
stack_preds = np.array(stack_preds)[sp.argsort(indexes_cv)]
with open(
"cache/models/%s/cv_preds/%s%d.pkl" %
(self.cache_dir, cache_file), 'wb') as f:
pickle.dump(stack_preds, f, pickle.HIGHEST_PROTOCOL)
return stack_preds
def _get_model_preds(self, model, X_train, X_predict, y_train, cache_file):
"""
Return the model predictions on the prediction set,
using cache if possible.
"""
model_output = load_from_cache(
"cache/models/%s/%s.pkl" % (self.cache_dir, cache_file),
self.use_cached_models)
model_params, model_preds = model_output \
if model_output is not None else (None, None)
if model_preds is None or model_params != model.get_params():
model.fit(X_train, y_train)
model_preds = model.predict_proba(X_predict)[:, 1]
with open("cache/models/%s/%s.pkl" % (
self.cache_dir, cache_file), 'wb') as f:
pickle.dump((model.get_params(), model_preds), f)
return model_preds
def _get_model_preds(self, model, X_train, X_predict, y_train, cache_file):
"""
Return the model predictions on the prediction set,
using cache if possible.
"""
model_output = load_from_cache(
"models/%s/%s.pkl" % (self.cache_dir, cache_file),
self.use_cached_models)
model_params, model_preds = model_output \
if model_output is not None else (None, None)
if model_preds is None or model_params != model.get_params():
model.fit(X_train, y_train)
model_preds = model.predict_proba(X_predict)[:, 1]
with open("cache/models/%s/%s.pkl" % (self.cache_dir, cache_file),
'wb') as f:
pickle.dump((model.get_params(), model_preds), f)
return model_preds
