Correlated Generators

[alecloudenback]

Introduce
closes #17

using EconomicScenarioGenerators, Copulas

m = BlackScholesMerton(0.01,0.02,.15,100.)
s = ScenarioGenerator(
                      1,  # timestep
                      30, # projection horizon
                      m,  # model
                  )

ss = [s,s] # these don't have to be the exact same
g = GaussianCopula([1. 0.9; 0.9 1.])
c = Correlated(ss,g)

collect(c)

using Plots
plot(collect(c))

cc @kasperrisager

Some disussion/todos:

• collect(c) returns a vector of vectors:

julia> collect(c)
2-element Vector{Vector{Float64}}:
 [87.58135277304397, 74.84820358945774, 68.56323172218646, 57.30284156605987, 58.01804063220005, 70.50583190735524, 67.62514257297967, 67.75331112863375, 74.65392532963543, 64.16076510394042  …  29.363724347621254, 39.05933584030768, 44.259288501108955, 30.193440520663124, 35.12786141576387, 32.4511890981983, 33.791351345428474, 36.62555404546912, 41.56345911563408, 37.00931733615785]
 [94.43341734287529, 82.88074095422284, 76.91159150462514, 72.64687376383395, 67.09645675465482, 70.74338844031544, 69.82729451232233, 64.10372892726669, 73.7458039859808, 65.07849904945566  …  27.412784359429537, 30.282329238312855, 29.048030126437794, 21.385593116622168, 27.28257459517675, 23.8838298202843, 23.561602356185862, 26.207618041390134, 29.221929707995933, 26.89015829202149]

• Only implemented for BlackScholesMerton currently, but could be extended to others pretty easily as long as the following are true (I'm not sure):
  • All models ultimately only need one random variate for the diffusion process. If there are multiple needed I'm not sure how that would be handled
  • Currently assumes that the random variate is gaussian - is there a way to pass the copula and use the marginal inverse CDF to get the variate for use in the diffusion process?
    • Maybe don't need to depend explicitly on Copulas given its interface?

[alecloudenback]

I posted about the vector of vector issues here: https://discourse.julialang.org/t/inverting-default-collect-behavior-when-iterator-returns-an-array/85409

[codecov]

Codecov Report

Merging #39 (d35aad1) into main (1d43cb9) will increase coverage by 4.06%.
The diff coverage is 100.00%.

@@            Coverage Diff             @@
##             main      #39      +/-   ##
==========================================
+ Coverage   90.16%   94.23%   +4.06%     
==========================================
  Files           4        4              
  Lines          61      104      +43     
==========================================
+ Hits           55       98      +43     
  Misses          6        6              

Impacted Files	Coverage Δ
src/EconomicScenarioGenerators.jl	96.49% <100.00%> (+4.49%)	⬆️
src/Yields.jl	87.50% <100.00%> (+9.72%)	⬆️
src/equity.jl	75.00% <100.00%> (+3.57%)	⬆️
src/interest.jl	100.00% <100.00%> (ø)

📣 Codecov can now indicate which changes are the most critical in Pull Requests. Learn more