将其编写为python代码

import numpy as np
from scipy.stats import norm, binom
from statsmodels.api import families, glm

n = 10000000
p = 10

x = np.random.normal(size=(n, p))
beta = np.arange(1, p+1).reshape(-1, 1)
z = x @ beta

condprob = norm.cdf(z)
y = binom.rvs(1, condprob, size=n).reshape(-1, 1)

prob_fit = glm(y, x, family=families.Binomial(link=families.links.probit)).fit()
logit_fit = glm(y, x, family=families.Binomial(link=families.links.logit)).fit()
linear_fit = glm(y, x, family=families.Gaussian(link=families.links.identity)).fit()

coef_mat = np.column_stack((prob_fit.params, logit_fit.params, linear_fit.params))
print(coef_mat)

prop_mat = np.column_stack((prob_fit.params / logit_fit.params,
                            prob_fit.params / linear_fit.params,
                            logit_fit.params / linear_fit.params))
print(prop_mat)

注：为了实现类似的功能，Python中需要导入一些额外的库，如scipy和statsmodels。