Open KapilKhanal opened 4 months ago
Possibly at some point but note you can compute these with auto-diff, here we compute the integral of $\exp(x)/(1.00001 - x)^2$:
julia> P = Legendre(); f = expand(P, exp); Sf = z -> stieltjes(f, z);
julia> @time ForwardDiff.derivative(Sf, 1.00001)
0.000015 seconds (9 allocations: 816 bytes)
-271798.40516450343
julia> @test ForwardDiff.derivative(Sf, 1.00001) ≈ -271798.4051645036 # mathematica w/ high precision
Test Passed
Awesome, I have the integrals 1/x^2 , 1/x^3. Does it work for these kind of kernels? I do not have test data point for these.
Also, I am confused a little bit on why autodiff would work in these cases. So if I have an algorithm for analytical integrals (that removes singularity somehow) of (1/x) , I can autodiff it to get the integrals of (1/x^2) and so on?
Hi I was just wondering if there are any plans for integrals of type
r^{-2 }
and above.Thanks,