如何匹配布尔索引数组与维度?
我想评估特征选择技术粒子群优化与神经网络。但我一直有这样的错误: “布尔索引与维度1上的索引数组不匹配;维数为1，对应的布尔维数为19。 这是我的代码:
model = Sequential()
model.add(Dense(16, input_dim=19, activation= 'relu'))
model.add(Dense(12, activation= 'relu'))
model.add(Dense(2, activation= 'softmax'))
model.add(Flatten())
model.compile(loss='sparse_categorical_crossentropy', optimizer='adam', metrics=['accuracy'])
def f_per_particle(m, alpha):
total_features = 19
if np.count_nonzero(m) == 0:
X_subset = x_train
else:
X_subset = x_train[:,m==1]
P=model.fit(X_subset, y_train, batch_size=64, epochs=100)
j = (alpha * (1.0 - P)+ (1.0 - alpha) * (1 - (X_subset.shape[1] / total_features)))
return j
def f(x, alpha=0.9):
n_particles = x.shape[0]
j = [f_per_particle(x[i], alpha) for i in range(n_particles)]
return np.array(j)
options = {'c1': 0.5, 'c2': 0.5, 'w':0.9, 'k': 30, 'p':2}
dimensions = 19
optimizer = ps.discrete.BinaryPSO(n_particles=100, dimensions=dimensions, options=options)
cost, pos = optimizer.optimize(f, iters=100)
问题来源StackOverflow 地址：/questions/59386403/how-to-match-boolean-index-array-along-with-dimension
写回答
-
After an update of Macports, that I think updated numpy, I'm getting the warning:
VisibleDeprecationWarning: boolean index did not match indexed array along dimension 1; dimension is 2 but corresponding boolean dimension is 1 inliers = n.size(pixels[distances <= self.dst]) that was not raised before. The related code is:
Compute distance of all non-zero points from the circumference
distances = guess_feature.points_distance(pixels)
Check which points are inliers (i.e. near the circle)
inliers = n.size(pixels[distances <= self.dst]) self.dst is a single scalar.
guess_feature.points_distance:
def points_distance(self,points): r''' Compute the distance of the points from the feature
:math:`d = \left| \sqrt{(x_i - x_c)^2 + (y_i-y_c)^2} - r \right|` Args: points (numpy.ndarray): a (n,2) numpy array, each row is a 2D Point. Returns: d (numpy.ndarray): the computed distances of the points from the feature. ''' xa = n.array([self.xc,self.yc]).reshape((1,2)) d = n.abs(dist.cdist(points,xa) - self.radius) return dAny ideas?
解决方案 I started getting a similar error after going up to numpy 1.10.1. I think you can get rid of the warning just by wrapping the boolean array in a numpy.where().
inliers = n.size(pixels[n.where(distances <= self.dst)]) Since you're just taking the size, there's no need to use the pixels array, so this should work:
inliers = n.size(n.where(distances <= self.dst])[0]) 机译 在Macports更新之后，我认为已经更新了numpy，我得到了警告：
VisibleDeprecationWarning：布尔值索引沿维度1与索引数组不匹配；维为2，但对应的布尔维为1 inliers = n.size（pixels [distances< = self.dst]）
以前没有提出过。相关代码为：
＃计算所有非零点到圆周的距离 的距离= guess_feature.points_distance （像素）
＃检查哪些点是inliers（即圆附近） inliers = n.size（pixels [distances <= self.dst]）
self.dst 是单个标量。
guess_feature.points_distance ：
def points_distance（self，points）： r''' 计算点距特征
的距离：math：
d = ＼ ＼左| \sqrt {（x_i-x_c）^ 2 +（y_i-y_c）^ 2}-r \right |Args： 分（numpy.ndarray）：a（ n，2）numpy数组，每一行都是一个2D点。
返回： d（numpy.ndarray）：点到要素的计算距离。
'''
xa = n.array（[self.xc，self.yc]）。reshape（（1,2）） d = n .abs（dist.cdist（points，xa）-self.radius） return d
2021-02-20 10:37:16赞同 展开评论 打赏
版权声明：本文内容由阿里云实名注册用户自发贡献，版权归原作者所有，阿里云开发者社区不拥有其著作权，亦不承担相应法律责任。具体规则请查看《阿里云开发者社区用户服务协议》和《阿里云开发者社区知识产权保护指引》。如果您发现本社区中有涉嫌抄袭的内容，填写侵权投诉表单进行举报，一经查实，本社区将立刻删除涉嫌侵权内容。
