SERVING THE QUANTITATIVE FINANCE COMMUNITY

 
User avatar
katastrofa
Posts: 9324
Joined: August 16th, 2007, 5:36 am
Location: Alpha Centauri

Re: How good or bad are the pandemic warning models?

April 28th, 2020, 6:33 pm

@"Dangerous; nettles and ticks."

Nettles are indeed risky. The roots contain a protein (UDA), which is superantigen (the immune system is strongly triggered by them as if it was attacked by pathogens) - further stimulation of the immune system doesn't sound like a good idea. I've found some nonsensical research on "mouse models" announcing that UDA could have protective effects against coronaviruses in humans... the animal-abusing psychos will come up with any theory to keep their business going.
BTW, it's curious whether other us (e.g. from wheat) alter the covid infection process in any way. Italy are pasta lovers...
I hope nobody takes seriously my medical aspirations ;-)

I remember once in Central America I once woke up at night with a realisation that I had a tick in an exact spot on my body. I obviously couldn't feel anything, because kininase in tick's saliva blocks pain and itching. It was there! A very tiny guy, not a cow-size

edit: yeah, lectins attach themselves to coronaviruses: https://www.tandfonline.com/doi/full/10 ... 10.4.2.291
But where does the other end go? Straight to the cell as they normally do to transport sugars - giving the virus a ride to the cell surface?
Just because they form a sort of an immune system for plants, it doesn't mean they will protect us - well, certainly not no-vegans.
I have no idea what I'm talking about, but I know that I sound very smart. I 'm thinking about starting a Twitter account...
Image
Last edited by katastrofa on April 28th, 2020, 7:26 pm, edited 1 time in total.
 
User avatar
Cuchulainn
Posts: 62387
Joined: July 16th, 2004, 7:38 am
Location: Amsterdam
Contact:

Re: How good or bad are the pandemic warning models?

April 28th, 2020, 7:00 pm

Is it not time to replace SIR by a Deep Learning approach?
 
User avatar
katastrofa
Posts: 9324
Joined: August 16th, 2007, 5:36 am
Location: Alpha Centauri

Re: How good or bad are the pandemic warning models?

April 28th, 2020, 7:27 pm

Yes, Deep Recurrent NNs if we won't find the vaccine before the next covid season.
 
User avatar
Cuchulainn
Posts: 62387
Joined: July 16th, 2004, 7:38 am
Location: Amsterdam
Contact:

Re: How good or bad are the pandemic warning models?

April 28th, 2020, 7:55 pm

Yes, Deep Recurrent NNs if we won't find the vaccine before the next covid season.
CTRNN?

https://neat-python.readthedocs.io/en/latest/ctrnn.html

Just retrofit Euler and GD and get it!

Lyapunov, anyone?
https://en.wikipedia.org/wiki/Lyapunov_stability
 
User avatar
katastrofa
Posts: 9324
Joined: August 16th, 2007, 5:36 am
Location: Alpha Centauri

Re: How good or bad are the pandemic warning models?

April 28th, 2020, 8:21 pm

Yes, Deep Recurrent NNs if we won't find the vaccine before the next covid season.
CTRNN?

https://neat-python.readthedocs.io/en/latest/ctrnn.html
It looks almost like differential equations. They are slowly getting there.
 
User avatar
Cuchulainn
Posts: 62387
Joined: July 16th, 2004, 7:38 am
Location: Amsterdam
Contact:

Re: How good or bad are the pandemic warning models?

April 29th, 2020, 11:27 am

Yes, Deep Recurrent NNs if we won't find the vaccine before the next covid season.
CTRNN?

https://neat-python.readthedocs.io/en/latest/ctrnn.html
It looks almost like differential equations. They are slowly getting there.
I could never understand why ML models are discrete up front?? Maybe CS education does not do ODE. It could be as simple as that..
 
User avatar
katastrofa
Posts: 9324
Joined: August 16th, 2007, 5:36 am
Location: Alpha Centauri

Re: How good or bad are the pandemic warning models?

April 29th, 2020, 8:51 pm

I'm guessing that the problems they modelled were usually simple enough to assume regular discrete time steps. In a large class of klooooooooooreal world systems (brain, any control systems, HFT, etc.) you often have long lags between events so it's ineffective to iterate RNN cell of a constant width  them and produces large errors.
 
User avatar
Cuchulainn
Posts: 62387
Joined: July 16th, 2004, 7:38 am
Location: Amsterdam
Contact:

Re: How good or bad are the pandemic warning models?

April 30th, 2020, 7:36 pm

I'm guessing that the problems they modelled were usually simple enough to assume regular discrete time steps. In a large class of klooooooooooreal world systems (brain, any control systems, HFT, etc.) you often have long lags between events so it's ineffective to iterate RNN cell of a constant width  them and produces large errors.
It's a form of colour-blindness.
 
User avatar
katastrofa
Posts: 9324
Joined: August 16th, 2007, 5:36 am
Location: Alpha Centauri

Re: How good or bad are the pandemic warning models?

April 30th, 2020, 8:39 pm

Colour-blindness? (Sorry for the strange typos, my cat likes to dance on my keyboard.)
 
User avatar
Cuchulainn
Posts: 62387
Joined: July 16th, 2004, 7:38 am
Location: Amsterdam
Contact:

Re: How good or bad are the pandemic warning models?

April 30th, 2020, 9:24 pm

Colour-blindness? (Sorry for the strange typos, my cat likes to dance on my keyboard.)
Not you, the others..
 
User avatar
Collector
Topic Author
Posts: 4647
Joined: August 21st, 2001, 12:37 pm

Re: How good or bad are the pandemic warning models?

May 1st, 2020, 8:12 am

how many more weeks before the number of people that labeled died of Covid-19 reach the number of people born in one day?
 
User avatar
katastrofa
Posts: 9324
Joined: August 16th, 2007, 5:36 am
Location: Alpha Centauri

Re: How good or bad are the pandemic warning models?

May 1st, 2020, 9:48 am

1,662,558 Deaths caused by smoking this year
831,804 Deaths caused by alcohol this year
559,082 Deaths caused by HIV/AIDS this year
448,943 Road traffic accident fatalities this year
356,637 Suicides this year
365,000 Approx average number of birth per day [UNICEF]
326,218 Deaths caused by malaria this year
234,393 Deaths caused by Covid-19
161,834 Seasonal flu deaths this year
102,795 Deaths of mothers during birth this year

Let's wait for the autumn season.
 
User avatar
Collector
Topic Author
Posts: 4647
Joined: August 21st, 2001, 12:37 pm

Re: How good or bad are the pandemic warning models?

May 1st, 2020, 11:23 am

yes the baby boom will start in the fall.
 
User avatar
katastrofa
Posts: 9324
Joined: August 16th, 2007, 5:36 am
Location: Alpha Centauri

Re: How good or bad are the pandemic warning models?

May 1st, 2020, 11:56 am

You keep posting about it - do you have some "anecdotal evidence" to announce? :-)
Or the whole statistics... :-O
 
User avatar
Paul
Posts: 10787
Joined: July 20th, 2001, 3:28 pm

Re: How good or bad are the pandemic warning models?

May 1st, 2020, 5:59 pm

yes the baby boom will start in the fall.
How many babies are the Chinese allowed to have these days?
ABOUT WILMOTT

PW by JB

Wilmott.com has been "Serving the Quantitative Finance Community" since 2001. Continued...


Twitter LinkedIn Instagram

JOBS BOARD

JOBS BOARD

Looking for a quant job, risk, algo trading,...? Browse jobs here...


GZIP: On