Use CovarianceEstimation.jl

[FriesischScott]

Using the package enables us to always estimate invertible covariance
matrices for the proposal distribution during the MCMC step.

I also extracted the bisection method estimating beta and the weights to
clean up the code.

[codecov-commenter]

Codecov Report

Merging #26 (68e601c) into main (b43348e) will decrease coverage by 1.16%.
The diff coverage is 95.23%.

@@             Coverage Diff             @@
##              main      #26      +/-   ##
===========================================
- Coverage   100.00%   98.83%   -1.17%     
===========================================
  Files            3        3              
  Lines           93       86       -7     
===========================================
- Hits            93       85       -8     
- Misses           0        1       +1     

Impacted Files	Coverage Δ
src/TransitionalMCMC.jl	100.00% <ø> (ø)
src/mcmc.jl	100.00% <ø> (ø)
src/tmcmc.jl	97.87% <95.23%> (-2.13%)	⬇️

---

Continue to review full report at Codecov.

Legend - Click here to learn more
Δ = absolute <relative> (impact), ø = not affected, ? = missing data
Powered by Codecov. Last update b43348e...68e601c. Read the comment docs.

[AnderGray]

Thanks for these changes, certainly an improvement and cleans up the code well.

Any idea how this new way of computing the covariance matrix compares to the previous way?

[AnderGray]

Looking good!