统计菜鸟求助，请高手帮忙看看可以给点建议或方法？万分感谢！
1. Suppose we are fitting y=K(a)e^at+K(b)e^bt where a and b are given. Can we rewrite K(a)e^at+K(b)e^bt in terms of K(a) and, say, Tb, so that k(a) and Tb can be separately estimated? How? Can we rewrite it so that K(b) and ,say, Ta an be separately estimated? How? What are the variances of K(a) after the first rewriting and of K(b) after the second? What were they in the beginning?
2. Suppose b is close enough to a so that e^(b-a)t=1+(b-a)t is a satisfactory approximation; what is the approximate variance of K(a)? What would happen if b were very close to a?
