本文整理汇总了Python中sklearn.preprocessing.Imputer.astype方法的典型用法代码示例。如果您正苦于以下问题：Python Imputer.astype方法的具体用法？Python Imputer.astype怎么用？Python Imputer.astype使用的例子？那么恭喜您, 这里精选的方法代码示例或许可以为您提供帮助。您也可以进一步了解该方法所在类sklearn.preprocessing.Imputer的用法示例。
在下文中一共展示了Imputer.astype方法的2个代码示例，这些例子默认根据受欢迎程度排序。您可以为喜欢或者感觉有用的代码点赞，您的评价将有助于系统推荐出更棒的Python代码示例。
示例1: affinity_propagation
# 需要导入模块: from sklearn.preprocessing import Imputer [as 别名]
# 或者: from sklearn.preprocessing.Imputer import astype [as 别名]
def affinity_propagation(location, location_callback):
"""Returns one or more clusters of a set of points, using an affinity
propagation algorithm.
The result is sorted with the first value being the largest cluster.
Returns:
A list of NamedTuples (see get_cluster_named_tuple for a definition
of the tuple).
"""
pts = location._tuple_points()
if not pts:
return None
X = np.array(pts).reshape((len(pts), len(pts[0])))
if np.any(np.isnan(X)) or not np.all(np.isfinite(X)):
return None
X = Imputer().fit_transform(X)
X = X.astype(np.float32)
afkwargs = {
'damping': 0.5,
'convergence_iter': 15,
'max_iter': 200,
'copy': True,
'preference': None,
'affinity': 'euclidean',
'verbose': False
}
af = AffinityPropagation(**afkwargs).fit(X)
cluster_centers_indices = af.cluster_centers_indices_
clusters = []
for cluster_id, cluster_centre in enumerate(af.cluster_centers_):
locations = []
for j, label in enumerate(af.labels_):
if not label == cluster_id:
continue
locations.append(location.locations[j])
if not locations:
continue
clusters.append(cluster_named_tuple()(label=cluster_id,
centroid=Point(cluster_centre),
location=location_callback(
locations)))
return clusters示例2: mean_shift
# 需要导入模块: from sklearn.preprocessing import Imputer [as 别名]
# 或者: from sklearn.preprocessing.Imputer import astype [as 别名]
def mean_shift(location, location_callback, bandwidth=None):
"""Returns one or more clusters of a set of points, using a mean shift
algorithm.
The result is sorted with the first value being the largest cluster.
Kwargs:
bandwidth (float): If bandwidth is None, a value is detected
automatically from the input using estimate_bandwidth.
Returns:
A list of NamedTuples (see get_cluster_named_tuple for a definition
of the tuple).
"""
pts = location._tuple_points()
if not pts:
return None
X = np.array(pts).reshape((len(pts), len(pts[0])))
if np.any(np.isnan(X)) or not np.all(np.isfinite(X)):
return None
X = Imputer().fit_transform(X)
X = X.astype(np.float32)
if not bandwidth:
bandwidth = estimate_bandwidth(X, quantile=0.3)
ms = MeanShift(bandwidth=bandwidth or None, bin_seeding=False).fit(X)
clusters = []
for cluster_id, cluster_centre in enumerate(ms.cluster_centers_):
locations = []
for j, label in enumerate(ms.labels_):
if not label == cluster_id:
continue
locations.append(location.locations[j])
if not locations:
continue
clusters.append(cluster_named_tuple()(label=cluster_id,
centroid=Point(cluster_centre),
location=location_callback(
locations)))
return clusters