find a way to solve the Ziggurat method # DeepLearning

[adamLutPhi]

[adamLutPhi]

Ziggurat algorithm is already proven to be biased in sample generation where samples tend to be thus generating unintended systematic bias in addition to the fact of it being Unable to Fairly Generate Samples in the Tail Area - which was the Actual Deal Breaker
#closed

[adamLutPhi]

Due to the fact that is has took a lot of Time & less actual Effort, in addition of being Surprised by the time taken doing it
shows a huge **Gap of Knowledge** in Algorithm Building, that's necessary to completely finish coding it
Beside, I even was surprised, the mass time wasted, not even in Procrastinating, but Avoiding to think, visualize or any sort of imagination about it - maybe just Incompatible Chemistry

New Repo?
I'd love to completely devote a _New Repo_ where special care would be provided, next time, when I am more Leveled-up, on the Technical level, where I would have gathered enough
It would be only focused on classical statistics (I took many times at least 3 occurrences since start of academic study) & its algorithms (or **Frequentist** algorithms (as viewed by a typical Bayesian)-whatever name you may call it

This Repo
I'll commence with it right after this , Big Guy;after starting on the Crux of this project, the one I should have been focused my lens on, in the first place:

1. Building an actual Neural Net, let's start with the adorable & Memorable McCulloch-Pitts Model, in Julia as a first time for me, that would be thrilling getting into it

2. I've actually started to working on its _helper structs - I'm not a left-brain 'I presume', I'm abit more a righty, & do trust my intuition. the only way to focus on the _real deal_ is by building around it, like a 3d helix (a whirlpool, so to speak) looking from above, starting from the sides, closing in on the target, slowly, - that is a Reverse Fibonacci sequence ( or prime numbers)

No need to rush things , lets put all the baggage out, near the doorway, & get my hands cleaned a bit )
funny thing, this should be monumental, but I'm not feeling it. On second thought, maybe it's better to feel this way, you know...

[adamLutPhi]

• Ziggurat function seems to be a Biased Data Sampler, where Samples are More around the Mean, whereas
  An Ideal Sampler would have Uniformly Samples spread across (on the x-Axis), regardless the Distribution & its respective properties
  Alas, Research on the Tail Area is to be quite ripe, for the time-being.
  (For this reason, I don't see it as an enhancement). Need to Cross-reference with another paper, about Tails , & proves this algorithm about (Fair Data-Sampling, around tails)