the definition of mgf:M(t)=E(e^tx)

if X1 and X2 are independent, then M(t1,t2)=E(e^t1x1+e^t2x2)



为何是 e^t1x1和e^t2x2相加的expectation呢?对mgf不适很理解,所以出现这种joint

mgf就更搞不

清楚了。谢谢
8 天 后
根据佚名统计学家的公式,就是随机变量函数的数学期望的公式得到的,矩母函数的分析性质比分布函数好,求矩和卷积要方便很多,当然它的缺陷在于不一定每个分布都存在矩母函数,而特征函数一定存在,
Since the mgf of X (random variable) is  

M(X)(t)=E(e^tX)=∫(e^tX)f(x)dx



Then , the joint mgf of X1 and X2 is

M(X1,X2)(t1,t2)=E(e^(t1*X1+t2*X2))=∫[e^(t1*X1+t2*X2)]f(x1,x2)dx1dx2



And if the X1 and X2 are independent, then

M(X1,X2)(t1,t2)=M(X1)(t1)*M(X2)(t2)