-
Neural networks make predictions based on data fed to them. Was there any data being used that was particularly interesting/novel? If it's too early to say anything about the impact of Omicron because there's not enough data that'll also be true for our (your) ability to assess a model - as training data will be incomplete - no?
You're bang on the money there. It'll likely be two or three months before there is enough data to be useful to the models. We were just setting up a few early so that we can keep an eye on predictions as the data improves.
With the Delta wave, it was about three months before the data matured enough to be useful.
My opinions aren't related to what the (currently useless) ML models show.
Stonehedge-
My opinions aren't related to what the (currently useless) ML models show.
Seemed they may be as you've mentioned it a couple of times and I couldn't work out what the contribution actually was.
Collating data sources and studying them to get a better idea how to use them and feed them into the models.
"Big data" meets "black box". (Not really having a go - this is what a large segment of research in my field does to some extent).
Neural networks make predictions based on data fed to them. Was there any data being used that was particularly interesting/novel? If it's too early to say anything about the impact of Omicron because there's not enough data that'll also be true for our (your) ability to assess a model - as training data will be incomplete - no?