From 10c146f1640cf66cf4dbcf79552a2ecf5d50ea55 Mon Sep 17 00:00:00 2001
From: Gael varoquaux <gael.varoquaux@normalesup.org>
Date: Mon, 19 Sep 2011 01:32:54 +0200
Subject: [PATCH] Cosmit: make in-place modifications explicit
---
.../grid_search_text_feature_extraction.py | 1 -
sklearn/cluster/k_means_.py | 21 ++++++++++++-------
2 files changed, 13 insertions(+), 9 deletions(-)
diff --git a/examples/grid_search_text_feature_extraction.py b/examples/grid_search_text_feature_extraction.py
index c89381afdb..d17c9f250c 100644
--- a/examples/grid_search_text_feature_extraction.py
+++ b/examples/grid_search_text_feature_extraction.py
@@ -49,7 +49,6 @@ print __doc__
from pprint import pprint
from time import time
-import os
import logging
from sklearn.datasets import fetch_20newsgroups
diff --git a/sklearn/cluster/k_means_.py b/sklearn/cluster/k_means_.py
index d709657cab..896da3eabc 100644
--- a/sklearn/cluster/k_means_.py
+++ b/sklearn/cluster/k_means_.py
@@ -575,11 +575,11 @@ def _mini_batch_step_dense(X, batch_slice, centers, counts, x_squared_norms):
The row slice of the mini batch.
centers: array, shape (k, n_features)
- The cluster centers
+ The cluster centers. This array is MODIFIED IN PLACE
counts: array, shape (k, )
The vector in which we keep track of the numbers of elements in a
- cluster
+ cluster. This array is MODIFIED IN PLACE
x_squared_norms: array, shape (n_samples,)
Squared euclidean norm of each data point.
@@ -599,6 +599,7 @@ def _mini_batch_step_dense(X, batch_slice, centers, counts, x_squared_norms):
centers[q] = (1. / (counts[q] + c)) * (
counts[q] * centers[q] + np.sum(X[center_mask], axis=0))
counts[q] += c
+ return counts, centers
def _mini_batch_step_sparse(X, batch_slice, centers, counts, x_squared_norms):
@@ -614,11 +615,11 @@ def _mini_batch_step_sparse(X, batch_slice, centers, counts, x_squared_norms):
The row slice of the mini batch.
centers: array, shape (k, n_features)
- The cluster centers
+ The cluster centers. This array is MODIFIED IN PLACE
counts: array, shape (k, )
The vector in which we keep track of the numbers of elements in a
- cluster
+ cluster. This array is MODIFIED IN PLACE
x_squared_norms: array, shape (n_samples,)
The squared norms of each sample in `X`.
@@ -628,6 +629,7 @@ def _mini_batch_step_sparse(X, batch_slice, centers, counts, x_squared_norms):
_k_means._mini_batch_update_sparse(X.data, X.indices, X.indptr,
batch_slice, centers, counts, cache)
+ return counts, centers
class MiniBatchKMeans(KMeans):
@@ -742,8 +744,10 @@ class MiniBatchKMeans(KMeans):
for i, batch_slice in izip(n_iterations, cycle(batch_slices)):
old_centers = self.cluster_centers_.copy()
- _mini_batch_step(X_shuffled, batch_slice, self.cluster_centers_,
- self.counts, x_squared_norms=x_squared_norms)
+ self.counts, self.cluster_centers_ = _mini_batch_step(
+ X_shuffled, batch_slice,
+ self.cluster_centers_, self.counts,
+ x_squared_norms=x_squared_norms)
if np.sum((old_centers - self.cluster_centers_) ** 2) < tol:
if self.verbose:
@@ -798,8 +802,9 @@ class MiniBatchKMeans(KMeans):
else:
_mini_batch_step = _mini_batch_step_dense
- _mini_batch_step(X, batch_slice, self.cluster_centers_, self.counts,
- x_squared_norms=x_squared_norms)
+ self.counts, self.cluster_centers_ = _mini_batch_step(X,
+ batch_slice, self.cluster_centers_, self.counts,
+ x_squared_norms=x_squared_norms)
self.inertia_, self.labels_ = _calculate_labels_inertia(
X, self.cluster_centers_, x_squared_norms)
--
GitLab