Hello,
I was just wondering if there was a more comprehensive function than nusum.
I have noticed that nusum fails to convert back many power series. Although
maybe this is on purpose as power series are not always guaranteed to
converge to the function they come from?
I'm looking for something like "unpowerseries" (perhaps using some simple
processing followed by a lookup table, or some method similar to how maxima
does integration).
The expected behavior I would like from such a function can be demonstrated
as:
(i1): unpowerseries( sum( exp(t*k) * lambda^(-k) * exp(-lambda) / k! )
(t1): exp(-lambda) * sum( (lambda*exp(t))^k / k! )
(t2): exp(-lambda) * sum( u^k / k! )
(t3): exp(-lambda) * exp(u)
(o1): exp(lambda * (exp(t) - 1).
It need not show the steps t1--t3, but I wanted to show how such a method
may operate.