@beadsland you're right is the alternative to TensorFlow, although there's also a TensorFlow version of most of the models. There are examples linked on the GitHub repo in the paper. And live examples at kaggle for most of the models and datasets.

@tavoglc Now see, if you'd lead with a live demo link that would have made a world of difference! Not finding "kaggle" in the PDF preprint either, BTW.

Sounds like this is all a *very* big idea and figuring out how to package it for easy dissemination & adoption by others ain't gonna be easy. It may all fit in your head, may all be readily visible to you, because is your brainchild, but for the rest of us, is proverbial elephant with us only grasping blindly.

@beadsland yeah is for people in general just try to get the word around. I think bulk of the transmission happens at daytime and things that aid the circadian rhythm could be used as prophylactic treatment in general. And a bunch of viruses will pop out of nowhere for at least 15 years.

@tavoglc Not sure your initial reply makes those findings clear. It reads as marketing hype for a new forecasting tool rather than a statement communicating what your analysis might tell us about transmission patterns.

Granted, what you're doing in this work is very dense. Communicating it in a digestible manner is gonna be a challenge. Yet communicating proactive use of information and proactive use of information tech are different tasks.

@beadsland yes that's why I started to write a substack trying to explain everything a little bit more in depth but I'm not sure if I'm making myself clear.
https://open.substack.com/pub/tavoglc/p/statistical-analysis-of-biological?utm_source=share&utm_medium=android&r=1yafgh
I think this is a somewhat complete index of all the work I've done regarding COVID. Code, examples, blog posts and another preprint.
https://github.com/TavoGLC/SARSCov2Solar
And it's exactly a marketing campaign, the phrase was used on a TensorFlow advertising from last year.

Thanks I think I'm going to make a web app or something like chat GPT, perhaps that brings a little bit more attention.

@tavoglc As someone who, perhaps in some way likewise, lives in the shadow of big, expansive indicial ideas, this sounds to be a familiar challenge.

That said, you realize that the vast majority of us have had less than zero exposure to TensorFlow advertising, and many hereabouts on fedi react negatively to even the whiff of being campaigned?

Perhaps a web app, as in interface that addresses some specific question (i.e., not like GPT), might prove more effective.