Python 两参数最大化函数
我试着找到两个参数的值，它们叫a e b，最大化函数fx，a，b，b>0。 我写道：Python 两参数最大化函数,python,python-2.7,scipy,minimize,Python,Python 2.7,Scipy,Minimize,我试着找到两个参数的值，它们叫a e b，最大化函数fx，a，b，b>0。 我写道： a=0.1 #start value for a b=150 #start value for b n=len(x) def f(y,a,b): c=sum([np.log(1-a/b*i) for i in y]) return -n*np.log(b)+(1-a/a)*c minimize(f,x,args=(a,b)) 其中x是包含我的数据的数组 我得到以下错误： RuntimeWarn
a=0.1 #start value for a
b=150 #start value for b
n=len(x)
def f(y,a,b):
c=sum([np.log(1-a/b*i) for i in y])
return -n*np.log(b)+(1-a/a)*c
minimize(f,x,args=(a,b))
RuntimeWarning: invalid value encountered in log
c=sum([np.log(1-a/b*i) for i in x])
C:\Python27\lib\site-packages\numpy\core\_methods.py:26: RuntimeWarning: invalid value encountered in reduce
return umr_maximum(a, axis, None, out, keepdims)
fun: nan
hess_inv: array([[1, 0, 0, ..., 0, 0, 0],
[0, 1, 0, ..., 0, 0, 0],
[0, 0, 1, ..., 0, 0, 0],
...,
[0, 0, 0, ..., 1, 0, 0],
[0, 0, 0, ..., 0, 1, 0],
[0, 0, 0, ..., 0, 0, 1]])
jac: array([nan, nan, nan, nan, nan, nan, nan, nan, nan, nan, nan, nan, nan,
nan, nan, nan, nan, nan, nan, nan, nan, nan, nan, nan, nan, nan,
nan, nan, nan, nan, nan, nan, nan, nan, nan, nan, nan, nan, nan,
nan, nan, nan, nan, nan, nan, nan, nan, nan, nan, nan, nan, nan,
nan, nan, nan, nan, nan, nan, nan, nan, nan, nan, nan, nan, nan,
nan, nan, nan, nan, nan, nan, nan, nan, nan, nan, nan, nan, nan,
nan, nan, nan, nan, nan, nan, nan, nan, nan, nan, nan, nan, nan,
nan, nan, nan, nan])
message: 'Desired error not necessarily achieved due to precision loss.'
nfev: 97
nit: 0
njev: 1
status: 2
success: False
有人能帮我吗？尝试使用的bounds属性设置boundarys。可能您的b在某一点上为负值。对于这个问题，由于您正在搜索a的最佳值，b您的目标应该重新定义如下：
def f(x,y):
a = x[0] # Parameter 1
b = x[1] # Parameter 2
c=sum([np.log(1-a/b*i) for i in y])
return -n*np.log(b)+(1-a/a)*c
minimize(f,x,args=(y))