Skip to content
Snippets Groups Projects
Commit 2fc5a381 authored by Luc Giffon's avatar Luc Giffon
Browse files

update nn_omp now takes a forest size to give its predictions + get coef method

parent 611e1481
Branches
No related tags found
1 merge request!24Resolve "non negative omp"
Show whitespace changes
Inline Side-by-side
code/bolsonaro/models/nn_omp.py
+ 14
− 86
View file @ 2fc5a381
......@@ -95,8 +95,9 @@ class NonNegativeOrthogonalMatchingPursuit:
self.lst_intermediate_solutions = lst_intermediate_solutions
self._set_intercept(T_offset, y_offset, T_scale)
def predict(self, X, idx_prediction=None):
if idx_prediction is not None:
def predict(self, X, forest_size=None):
if forest_size is not None:
idx_prediction = self.requested_intermediate_solutions_sizes.index(forest_size)
return X @ self.lst_intermediate_solutions[idx_prediction] + self.lst_intercept[idx_prediction]
else:
predictions = []
......@@ -104,89 +105,16 @@ class NonNegativeOrthogonalMatchingPursuit:
predictions.append(X @ sol + self.lst_intercept[idx_sol])
return predictions
def nn_omp(T, y, max_iter, intermediate_solutions_sizes=None, force_return_all_solutions=True, logger=None):
def get_coef(self, forest_size=None):
"""
Ref: Sparse Non-Negative Solution of a Linear System of Equations is Unique
T: (N x L)
y: (N x 1)
max_iter: the max number of iteration. If requested_intermediate_solutions_sizes is None. Return the max_iter-sparse solution.
requested_intermediate_solutions_sizes: a list of the other returned intermediate solutions than with max_iter (they are returned in a list with same indexes)
return the intermediate solution corresponding to requested forest size if not None.
Return the list of intermediate solutions. If the perfect solution is found before the end, the list may not be full.
Else return the list of intermediate solution.
:param forest_size:
:return:
"""
if intermediate_solutions_sizes is None:
intermediate_solutions_sizes = [max_iter]
assert all(type(elm) == int for elm in intermediate_solutions_sizes), "All intermediate solution must be size specified as integers."
iter_intermediate_solutions_sizes = iter(intermediate_solutions_sizes)
lst_intermediate_solutions = []
bool_arr_selected_indexes = np.zeros(T.shape[1], dtype=bool)
residual = y
i = 0
next_solution = next(iter_intermediate_solutions_sizes, None)
while i < max_iter and next_solution != None and not np.isclose(np.linalg.norm(residual), 0):
# if logger is not None: logger.debug("iter {}".format(i))
# compute all correlations between atoms and residual
dot_products = T.T @ residual
idx_max_dot_product = np.argmax(dot_products)
# only positively correlated results can be taken
if dot_products[idx_max_dot_product] <= 0:
logger.warning("No other atoms is positively correlated with the residual. End prematurely with {} atoms.".format(i+1))
break
# selection of atom with max correlation with residual
bool_arr_selected_indexes[idx_max_dot_product] = True
tmp_T = T[:, bool_arr_selected_indexes]
sol = nnls(tmp_T, y)[0] # non negative least square
residual = y - tmp_T @ sol
if i+1 == next_solution:
final_vec = np.zeros(T.shape[1])
final_vec[bool_arr_selected_indexes] = sol # solution is full of zero but on selected indices
lst_intermediate_solutions.append(final_vec)
next_solution = next(iter_intermediate_solutions_sizes, None)
i+=1
nb_missing_solutions = len(intermediate_solutions_sizes) - len(lst_intermediate_solutions)
if len(lst_intermediate_solutions) == 1:
return lst_intermediate_solutions[-1]
if nb_missing_solutions > 0:
if force_return_all_solutions:
logger.warning("nn_omp ended prematurely and found less solution than expected: "
"expected {}. found {}".format(len(intermediate_solutions_sizes), len(lst_intermediate_solutions)))
return lst_intermediate_solutions.extend([deepcopy(lst_intermediate_solutions[-1]) for _ in range(len(intermediate_solutions_sizes) - len(lst_intermediate_solutions))])
else:
return lst_intermediate_solutions
if forest_size is not None:
idx_prediction = self.requested_intermediate_solutions_sizes.index(forest_size)
return self.lst_intermediate_solutions[idx_prediction]
else:
return lst_intermediate_solutions
if __name__ == "__main__":
N = 1000
L = 100
K = 10
T = np.random.rand(N, L)
w_star = np.abs(np.random.rand(L))
T /= np.linalg.norm(T, axis=0)
y = T @ w_star
requested_solutions = list(range(1, L, 10))
solutions = nn_omp(T, y, L, requested_solutions)
for idx_sol, w in enumerate(solutions):
solution = T @ w
non_zero = w.astype(bool)
print(requested_solutions[idx_sol], np.sum(non_zero), np.linalg.norm(solution - y)/np.linalg.norm(y))
return self.lst_intermediate_solutions
\ No newline at end of file
code/bolsonaro/models/nn_omp_forest_classifier.py
+ 13
− 13
View file @ 2fc5a381
......@@ -13,10 +13,10 @@ import warnings
class NonNegativeOmpForestBinaryClassifier(OmpForestBinaryClassifier):
def predict(self, X, idx_prediction=None):
def predict(self, X, forest_size=None):
"""
Make prediction.
If idx_prediction is None return the list of predictions of all intermediate solutions
If forest_size is None return the list of predictions of all intermediate solutions
:param X:
:return:
......@@ -26,30 +26,30 @@ class NonNegativeOmpForestBinaryClassifier(OmpForestBinaryClassifier):
if self._models_parameters.normalize_D:
forest_predictions /= self._forest_norms
return self._omp.predict(forest_predictions, idx_prediction)
return self._omp.predict(forest_predictions, forest_size)
def predict_no_weights(self, X, idx_prediction=None):
def predict_no_weights(self, X, forest_size=None):
"""
Make a prediction of the selected trees but without weight.
If idx_prediction is None return the list of unweighted predictions of all intermediate solutions.
If forest_size is None return the list of unweighted predictions of all intermediate solutions.
:param X: some data to apply the forest to
:return: a np.array of the predictions of the trees selected by OMP without applying the weight
"""
forest_predictions = np.array([tree.predict(X) for tree in self._base_forest_estimator.estimators_])
if idx_prediction is not None:
weights = self._omp.lst_intermediate_solutions[idx_prediction]
if forest_size is not None:
weights = self._omp.get_coef(forest_size)
select_trees = np.mean(forest_predictions[weights != 0], axis=0)
return select_trees
else:
lst_predictions = []
for sol in self._omp.lst_intermediate_solutions:
for sol in self._omp.get_coef():
lst_predictions.append(np.mean(forest_predictions[sol != 0], axis=0))
return lst_predictions
def score(self, X, y, idx_prediction=None):
def score(self, X, y, forest_size=None):
"""
Evaluate OMPForestClassifer on (`X`, `y`).
......@@ -60,8 +60,8 @@ class NonNegativeOmpForestBinaryClassifier(OmpForestBinaryClassifier):
:return:
"""
# raise NotImplementedError("Function not verified")
if idx_prediction is not None:
predictions = self.predict(X, idx_prediction)
if forest_size is not None:
predictions = self.predict(X, forest_size)
# not sure predictions are -1/+1 so might be zero percent accuracy
return np.sum(predictions != y) / len(y)
else:
......@@ -82,7 +82,7 @@ if __name__ == "__main__":
X, y, test_size = 0.33, random_state = 42)
# intermediate_solutions = [100, 200, 300, 400, 500, 1000]
intermediate_solutions = [10, 20, 30, 40, 50, 100]
intermediate_solutions = [10, 20, 30, 40, 50, 100, 300]
nnmodel_params = ModelParameters(extracted_forest_size=50,
normalize_D=True,
subsets_used=["train", "dev"],
......@@ -99,7 +99,7 @@ if __name__ == "__main__":
intermediate_solutions_sizes=intermediate_solutions)
extracted_size = 50
extracted_size = 300
nn_ompforest = NonNegativeOmpForestBinaryClassifier(nnmodel_params)
nn_ompforest.fit(X_train, y_train, X_train, y_train)
model_params = ModelParameters(extracted_forest_size=extracted_size,
......
code/bolsonaro/models/nn_omp_forest_regressor.py
+ 16
− 16
View file @ 2fc5a381
......@@ -12,10 +12,10 @@ from bolsonaro.models.omp_forest_regressor import OmpForestRegressor
class NonNegativeOmpForestRegressor(OmpForestRegressor):
def predict(self, X, idx_prediction=None):
def predict(self, X, forest_size=None):
"""
Make prediction.
If idx_prediction is None return the list of predictions of all intermediate solutions
If forest_size is None return the list of predictions of all intermediate solutions
:param X:
:return:
......@@ -25,30 +25,30 @@ class NonNegativeOmpForestRegressor(OmpForestRegressor):
if self._models_parameters.normalize_D:
forest_predictions /= self._forest_norms
return self._omp.predict(forest_predictions, idx_prediction)
return self._omp.predict(forest_predictions, forest_size)
def predict_no_weights(self, X, idx_prediction=None):
def predict_no_weights(self, X, forest_size=None):
"""
Make a prediction of the selected trees but without weight.
If idx_prediction is None return the list of unweighted predictions of all intermediate solutions.
If forest_size is None return the list of unweighted predictions of all intermediate solutions.
:param X: some data to apply the forest to
:return: a np.array of the predictions of the trees selected by OMP without applying the weight
"""
forest_predictions = np.array([tree.predict(X) for tree in self._base_forest_estimator.estimators_])
if idx_prediction is not None:
weights = self._omp.lst_intermediate_solutions[idx_prediction]
if forest_size is not None:
weights = self._omp.get_coef(forest_size)
select_trees = np.mean(forest_predictions[weights != 0], axis=0)
return select_trees
else:
lst_predictions = []
for sol in self._omp.lst_intermediate_solutions:
for sol in self._omp.get_coef():
lst_predictions.append(np.mean(forest_predictions[sol != 0], axis=0))
return lst_predictions
def score(self, X, y, idx_prediction=None):
def score(self, X, y, forest_size=None):
"""
Evaluate OMPForestClassifer on (`X`, `y`).
......@@ -59,8 +59,8 @@ class NonNegativeOmpForestRegressor(OmpForestRegressor):
:return:
"""
# raise NotImplementedError("Function not verified")
if idx_prediction is not None:
predictions = self.predict(X, idx_prediction)
if forest_size is not None:
predictions = self.predict(X, forest_size)
# not sure predictions are -1/+1 so might be zero percent accuracy
return np.mean(np.square(predictions - y))
else:
......@@ -71,13 +71,13 @@ class NonNegativeOmpForestRegressor(OmpForestRegressor):
return lst_scores
if __name__ == "__main__":
# X, y = load_boston(return_X_y=True)
X, y = fetch_california_housing(return_X_y=True)
X, y = load_boston(return_X_y=True)
# X, y = fetch_california_housing(return_X_y=True)
X_train, X_test, y_train, y_test = train_test_split(
X, y, test_size = 0.33, random_state = 42)
intermediate_solutions = [100, 200, 300, 400, 500, 1000]
nnmodel_params = ModelParameters(extracted_forest_size=600,
intermediate_solutions = [10, 20, 30, 40, 50, 100, 200]
nnmodel_params = ModelParameters(extracted_forest_size=60,
normalize_D=True,
subsets_used=["train", "dev"],
normalize_weights=False,
......@@ -95,7 +95,7 @@ if __name__ == "__main__":
nn_ompforest = NonNegativeOmpForestRegressor(nnmodel_params)
nn_ompforest.fit(X_train, y_train, X_train, y_train)
model_params = ModelParameters(extracted_forest_size=50,
model_params = ModelParameters(extracted_forest_size=200,
normalize_D=True,
subsets_used=["train", "dev"],
normalize_weights=False,
......
0% or .
You are about to add 0 people to the discussion. Proceed with caution.
Please register or to comment