[SystemSafety] Accuracy of COVID modeling code

[Mike Ellims]

An mildly interesting comparison of a current set of models given comments by Thierry previously; i.e. 

>> The modelling itself will lend to variability in outputs depending on the quality of the random
>> generator. The input parameters are only estimates, which can vary widely depending on the
>> assumptions on a disease which is not well known.
>> Any particular result should be taken with a lot of caution, and even the "average"
>> behavior on what is essentially an exponential model may be suspect.

In general this kind of model may be a case of measure with a micrometer, mark with chalk and cut with an axe.

The link: https://projects.fivethirtyeight.com/covid-forecasts/?ex_cid=rrpromo

Cheers.

-----Original Message-----
From: systemsafety [mailto:systemsafety-bounces at lists.techfak.uni-bielefeld.de] On Behalf Of Peter Bernard Ladkin
Sent: 06 May 2020 19:29
To: systemsafety at lists.techfak.uni-bielefeld.de
Subject: Re: [SystemSafety] Accuracy of COVID modeling code



On 2020-05-06 20:17 , Derek M Jones wrote:
> 
>> You mean R, compared with SAS, NAG, Mathematica, Matlab, Maple and so 
>> on? As I said, enormous amount of literature.
> 
> I mean any, as in any computer language, past or present.
> 
> There is certainly an enormous amount of literature containing claims 
> based on ego and bluster.
> 
> My statement involved the evidence-based literature, which is sparse 
> to say the least.
No. It is sufficiently voluminous to completely overwhelm the ability of any one expert to read, let alone to master, except for the occasional genius. As, indeed, nowadays in any well-developed area of mathematics.

Have you checked out the ACM Transactions on Mathematical Software, for example?

PBL

Prof. Peter Bernard Ladkin, Bielefeld, Germany Styelfy Bleibgsnd
Tel+msg +49 (0)521 880 7319  www.rvs-bi.de