baohan
/
scikit-learn
mirror of https://github.com/scikit-learn/scikit-learn.git
0
0
Fork 0
Code Issues Packages Projects Releases Wiki Activity

Change naming scheme of variables for consistency

This commit is contained in:
Johannes Schönberger 2013-09-27 10:36:54 +02:00 committed by Olivier Grisel
parent 018a947bb1
commit 69113d1663
1 changed files with 32 additions and 31 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

63
sklearn/linear_model/ransac.py
View File

@ -174,11 +174,11 @@ class RANSAC(BaseEstimator, MetaEstimatorMixin):
except ValueError:
pass
best_n_inliers = 0
best_score = np.inf
best_inlier_mask = None
best_inlier_X = None
best_inlier_y = None
n_inliers_best = 0
score_best = np.inf
inlier_mask_best = None
X_inlier_best = None
y_inlier_best = None
# number of data samples
n_samples = X.shape[0]
@ -194,62 +194,63 @@ class RANSAC(BaseEstimator, MetaEstimatorMixin):
# choose random sample set
random_idxs = random_state.randint(0, n_samples, min_n_samples)
rsample_X = X[random_idxs]
rsample_y = y[random_idxs]
X_subset = X[random_idxs]
y_subset = y[random_idxs]
# check if random sample set is valid
if (self.is_data_valid is not None
and not self.is_data_valid(rsample_X, rsample_y)):
and not self.is_data_valid(X_subset, y_subset)):
continue
# fit model for current random sample set
base_estimator.fit(rsample_X, rsample_y)
base_estimator.fit(X_subset, y_subset)
# check if estimated model is valid
if (self.is_model_valid is not None and not
self.is_model_valid(base_estimator, rsample_X, rsample_y)):
self.is_model_valid(base_estimator, X_subset,
y_subset)):
continue
# residuals of all data for current random sample model
rsample_residuals = residual_metric(base_estimator.predict(X) - y)
residuals_subset = residual_metric(base_estimator.predict(X) - y)
# classify data into inliers and outliers
rsample_inlier_mask = rsample_residuals < residual_threshold
rsample_n_inliers = np.sum(rsample_inlier_mask)
inlier_mask_subset = residuals_subset < residual_threshold
n_inliers_subset = np.sum(inlier_mask_subset)
# less inliers -> skip current random sample
if rsample_n_inliers < best_n_inliers:
if n_inliers_subset < n_inliers_best:
continue
# extract inlier data set
rsample_inlier_idxs = sample_idxs[rsample_inlier_mask]
rsample_inlier_X = X[rsample_inlier_idxs]
rsample_inlier_y = y[rsample_inlier_idxs]
inlier_idxs_subset = sample_idxs[inlier_mask_subset]
X_inlier_subset = X[inlier_idxs_subset]
y_inlier_subset = y[inlier_idxs_subset]
# score of inlier data set
rsample_score = base_estimator.score(rsample_inlier_X,
rsample_inlier_y)
score_subset = base_estimator.score(X_inlier_subset,
y_inlier_subset)
# same number of inliers but worse score -> skip current random
# sample
if (rsample_n_inliers == best_n_inliers
and rsample_score < best_score):
if (n_inliers_subset == n_inliers_best
and score_subset < score_best):
continue
# save current random sample as best sample
best_n_inliers = rsample_n_inliers
best_score = rsample_score
best_inlier_mask = rsample_inlier_mask
best_inlier_X = rsample_inlier_X
best_inlier_y = rsample_inlier_y
n_inliers_best = n_inliers_subset
score_best = score_subset
inlier_mask_best = inlier_mask_subset
X_inlier_best = X_inlier_subset
y_inlier_best = y_inlier_subset
# break if sufficient number of inliers or score is reached
if (best_n_inliers >= self.stop_n_inliers
or best_score >= self.stop_score):
if (n_inliers_best >= self.stop_n_inliers
or score_best >= self.stop_score):
break
# if none of the iterations met the required criteria
if best_inlier_mask is None:
if inlier_mask_best is None:
raise ValueError("RANSAC could not find valid consensus set, "
"because `is_data_valid` and `is_model_valid` "
"returned False for all `max_trials` randomly "
@ -257,10 +258,10 @@ class RANSAC(BaseEstimator, MetaEstimatorMixin):
"constraints.")
# estimate final model using all inliers
base_estimator.fit(best_inlier_X, best_inlier_y)
base_estimator.fit(X_inlier_best, y_inlier_best)
self.estimator_ = base_estimator
self.inlier_mask_ = best_inlier_mask
self.inlier_mask_ = inlier_mask_best
def predict(self, X):
"""Predict using the estimated model.