我有一个2d数组x,每行有不同数量的nan值:
array([[ nan, -0.355, -0.036, ..., nan, nan],
[ nan, -0.341, -0.047, ..., nan, 0.654],
[ .016, -1.147, -0.667, ..., nan, nan],
...,
[ nan, 0.294, -0.235, ..., 0.65, nan]])
给定此数组,对于每一行,我想计算前25个百分点内所有值的平均值。我正在做以下事情:
limit = np.nanpercentile(x, 25, axis=1) # output 1D array
ans = np.nanmean(x * (x < limit[:,None]), axis=1)
但是这给出了错误的结果 - 特别是计数(np.nansum / np.nanmean)无论我选择什么百分位数都保持不变,因为比较在不正确的地方产生零,并被计为有效值意思。我不能简单地使用x[x>limit[:,None]]
,因为它提供了一维数组,我需要一个2D结果。
我通过以下方式解决了这个问题:
f = x.copy()
f[f > limit[:,None]] = np.nan
ans = np.nanmean(f, axis=1)
有更好的方法吗?
答案 0 :(得分:2)
方法#1:您可以创建一个无效的掩码,它们是原始数组的NaNs
和f > limit[:,None]
的掩码。然后,使用此掩码执行np.nanmean
等效方法,只考虑masking
的有效方法。使用masks/boolean arrays
的好处在于内存,因为它占用的内存比浮动pt数组少8倍。因此,我们会有这样的实现 -
# Create mask of non-NaNs and thresholded ones
mask = ~np.isnan(x) & (x <= limit[:,None])
# Get the row, col indices. Use the row indices for bin-based summing and
# finally averaging by using those indices to get the group lengths.
r,c = np.where(mask)
out = np.bincount(r,x[mask])/np.bincount(r)
方法#2:我们也可以使用np.add.reduceat
这里的帮助,因为已经按照掩码对分类进行了分类。所以,效率会更高 -
# Get the valid mask as before
mask = ~np.isnan(x) & (x <= limit[:,None])
# Get valid row count. Use np.add.reduceat to perform grouped summations
# at intervals separated by row indices.
rowc = mask.sum(1)
out = np.add.reduceat(x[mask],np.append(0,rowc[:-1].cumsum()))/rowc
<强>基准强>
功能定义 -
def original_app(x, limit):
f = x.copy()
f[f > limit[:,None]] = np.nan
ans = np.nanmean(f, axis=1)
return ans
def proposed1_app(x, limit):
mask = ~np.isnan(x) & (x <= limit[:,None])
r,c = np.where(mask)
out = np.bincount(r,x[mask])/np.bincount(r)
return out
def proposed2_app(x, limit):
mask = ~np.isnan(x) & (x <= limit[:,None])
rowc = mask.sum(1)
out = np.add.reduceat(x[mask],np.append(0,rowc[:-1].cumsum()))/rowc
return out
计时和验证 -
In [402]: # Setup inputs
...: x = np.random.randn(400,500)
...: x.ravel()[np.random.randint(0,x.size,x.size//4)] = np.nan # Half as NaNs
...: limit = np.nanpercentile(x, 25, axis=1)
...:
In [403]: np.allclose(original_app(x, limit),proposed1_app(x, limit))
Out[403]: True
In [404]: np.allclose(original_app(x, limit),proposed2_app(x, limit))
Out[404]: True
In [405]: %timeit original_app(x, limit)
100 loops, best of 3: 5 ms per loop
In [406]: %timeit proposed1_app(x, limit)
100 loops, best of 3: 4.02 ms per loop
In [407]: %timeit proposed2_app(x, limit)
100 loops, best of 3: 2.18 ms per loop