关于“python和matlab怎么混合编程”这个问题，昨晚回答的比较仓促，抱歉。我搜了一下，你看看下面的几个方案是否能解决：

1：An Open-Source MATLAB®-to-Python® Compiler

http://ompc.juricap.com/

OMPC allows running MATLAB®’s m-files using Python interpreter. OMPC reads m-files and translates them into Python compatible code. Just like in the following example

从描述来看OMPC的目标是将Matlab代码转换成Python代码，可以试试利用OMPC“翻译”你的Matlab代码。

2. PyMat – An interface between Python and MATLAB

http://claymore.engineer.gvsu.edu/~steriana/Python/pymat.html

The PyMat module acts as an interface between NumPy arrays in Python and a MATLAB engine session, allowing arrays to be passed back and forth and arbitrary commands to be executed in the MATLAB workspace.

这个我估计你可能测试过了。

如果上述方案不能解决，这里再写个SOS吧，呵呵。

def Likelihoood(A,X=dataX,Y=dataY):
result=0

for i in range(0,len(X)):
tmp=F(A,X[i])
if tmp==1:
tmp==0.9999999
if tmp<=0:
tmp=0.0000001
print tmp
result=result+ math.log((1-tmp),e)+Y[i]*math.log((tmp/(1-tmp)),e)
return result
def GradientLikelihoood(A,X=dataX,Y=dataY,Z=dataZ):
#Z为正例个数
#求F梯度
result0=0
result1=0
gradientMatrix=zeros(A.shape)
for i in range(0,len(X)):
result0=result0+Y[i]-Z[i]*F(A,X[i])
result1=result1+(Y[i]-Z[i]*F(A,X[i]))*X[i]

gradientMatrix[0]=result0
gradientMatrix[1]=result1
return gradientMatrix
def HessianLikelihoood(A,X=dataX,Y=dataY,Z=dataZ):
HessianMatrix=zeros((2,2))
for i in range(0,len(X)):
tmp=F(A,X[i])*(1-F(A,X[i]))
HessianMatrix[0][0]=HessianMatrix[0][0]+Z[i]*tmp
HessianMatrix[0][1]=HessianMatrix[0][1]+X[i]*Z[i]*tmp
HessianMatrix[1][1]=HessianMatrix[1][1]=X[i]**2*Z[i]*tmp
HessianMatrix[1][0]=HessianMatrix[0][1]
return HessianMatrix

A0=[10,-9.6]
xopt = fmin_ncg(Likelihoood,A0,GradientLikelihoood,HessianLikelihoood)
print xopt