我试图定义一个包含内部循环的函数来模拟积分.
问题是速度.在我的机器上评估该功能一次最多可能需要30秒.由于我的最终目标是最小化这个功能,一些额外的速度会很好.
因此,我试图让Cython工作,但我必须犯一个严重的错误(可能很多人!).在Cython文档之后,我尝试输入我的变量.这样做之后,代码就像纯Python一样慢.这看起来很奇怪.
这是我的代码:
import numpy as np
cimport cython
cimport numpy as np
import minuit
data = np.genfromtxt('q6data.csv', usecols = np.arange(1, 24, 1), delimiter = ',')
cdef int ns = 1000 # Number of simulation draws
cdef int K = 5 # Number of observed characteristics, including constant
cdef int J = len(data[:,1]) # Number of products, including outside
cdef double tol = 0.0001 # Inner GMM loop tolerance
nu = np.random.normal(0, 1, (6, ns)) # ns random deviates …我已经定义了以下功能:
def GMM(s1, s2, s3, s4, s5, a):
"""The GMM objective function.
Arguments
---------
si: float
standard deviations of preference distribution
a: float
marginal utility of residutal income
Paramters
---------
Px: array (1,ns)
projector onto nonprice characteristic space
xk, z: arrays (J, 5) and (J, 12)
nonprice char. and instruments
invW: array (12, 12)
GMM weight matrix
Returns
-------
float."""
delta = invert(s1, s2, s3, s4, s5, a, delta0) # Invert market shares to get mean utility
bmean = np.dot(Px, …我正在尝试使用Cython来加速一段代码.当我使用lambda函数时,Cython会给出一个错误,上面写着"Expected a identifier或literal".据我所知,lambda函数在Cython 0.13中得到支持.我在这一点上不正确吗？如果它们确实受到支持,我是否需要做一些其他事情而不是我在这里实施它们？
def f(e_1, e_2, rho):
"""Bivariate Normal pdf with mean zero, unit variances, and correlation coefficient rho."""
return (1.0 / (2.0 * pi * sqrt(1 - rho**2))) * exp(-(1.0 / (2*(1 - rho**2))) * (e_1**2 + e_2**2 - 2*rho*e_1*e_2))
def P_zero(b_10, b_11, b_20, b_21, rho, gamma, x):
"""Returns the probability of observing zero entrants in a market by numerically
integrating out the unobserved firm-specific profit shocks."""
h_z = lambda e_1: -inf
g_z = lambda e_1: -b_10 - …