我有一个1-D函数需要花费很多时间来计算一个大的2-D'x'值数组,因此使用SciPy工具创建插值函数然后使用它来计算y很容易,哪个会快得多。但是,我不能在具有1-D以上的数组上使用插值函数。
示例:
# First, I create the interpolation function in the domain I want to work
x = np.arange(1, 100, 0.1)
f = exp(x) # a complicated function
f_int = sp.interpolate.InterpolatedUnivariateSpline(x, f, k=2)
# Now, in the code I do that
x = [[13, ..., 1], [99, ..., 45], [33, ..., 98] ..., [15, ..., 65]]
y = f_int(x)
# Which I want that it returns y = [[f_int(13), ..., f_int(1)], ..., [f_int(15), ..., f_int(65)]]
但回归:
ValueError: object too deep for desired array
我知道我可以遍历所有x成员,但我不知道这是否是更好的选择......
谢谢!
编辑:
这样的功能也可以完成这项工作:
def vector_op(function, values):
orig_shape = values.shape
values = np.reshape(values, values.size)
return np.reshape(function(values), orig_shape)
我已经尝试过np.vectorize,但它太慢了......
答案 0 :(得分:2)
如果f_int
需要单维数据,则应将输入展平,将其输入内插器,然后重建原始形状:
>>> x = np.arange(1, 100, 0.1)
>>> f = 2 * x # a simple function to see the results are good
>>> f_int = scipy.interpolate.InterpolatedUnivariateSpline(x, f, k=2)
>>> x = np.arange(25).reshape(5, 5) + 1
>>> x
array([[ 1, 2, 3, 4, 5],
[ 6, 7, 8, 9, 10],
[11, 12, 13, 14, 15],
[16, 17, 18, 19, 20],
[21, 22, 23, 24, 25]])
>>> x_int = f_int(x.reshape(-1)).reshape(x.shape)
>>> x_int
array([[ 2., 4., 6., 8., 10.],
[ 12., 14., 16., 18., 20.],
[ 22., 24., 26., 28., 30.],
[ 32., 34., 36., 38., 40.],
[ 42., 44., 46., 48., 50.]])
x.reshape(-1)
执行展平,.reshape(x.shape)
将其恢复为原始形式。
答案 1 :(得分:1)
我想你想在numpy中做一个矢量化函数:
#create some random test data
test = numpy.random.random((100,100))
#a normal python function that you want to apply
def myFunc(i):
return np.exp(i)
#now vectorize the function so that it will work on numpy arrays
myVecFunc = np.vectorize(myFunc)
result = myVecFunc(test)
答案 2 :(得分:0)
我会使用list comprehension
和map
的组合(可能有一种方法可以使用我缺少的两个嵌套maps
)
In [24]: x
Out[24]: [[1, 2, 3], [1, 2, 3], [1, 2, 3]]
In [25]: [map(lambda a: a*0.1, x_val) for x_val in x]
Out[25]:
[[0.1, 0.2, 0.30000000000000004],
[0.1, 0.2, 0.30000000000000004],
[0.1, 0.2, 0.30000000000000004]]
这仅用于说明目的....将lambda a: a*0.1
替换为您的函数f_int