Omega Owners Forum
Chat Area => General Discussion Area => Topic started by: STEMO on 03 February 2020, 12:03:59
-
My lad just phoned me to say he was disappointed at an exam he took at uni this morning. It was about Bayesian statistics. He reckoned he might have scraped it, but it was a knock to his confidence, as he thought he'd be ok.
I've never heard of this subject before, so thought I'd take a look online to see how much of it I understood.
Answer......none....none at all.....not even a tiny little bit ;D ;D
http://www.scholarpedia.org/article/Bayesian_statistics
-
Bayesian statistics is a system for describing epistemological uncertainty using the mathematical language of probability. In the 'Bayesian paradigm,' degrees of belief in states of nature are specified; these are non-negative, and the total belief in all states of nature is fixed to be one. Bayesian statistical methods start with existing 'prior' beliefs, and update these using data to give 'posterior' beliefs, which may be used as the basis for inferential decisions.
They lost me at 'describing'
-
You have a bright lad even for taking that exam.....total bo***cks as far I can understand from the description
-
Taking that is an achievement in its own right, and if he has any grasp of the concepts, then I am sure he will be pleasantly surprised with the result.
Regardless of how someone's mind works he cannot alter the result after the event and should already be looking forward to the next thing. Easier said than done, even for simple people like us, so I don't relish you having to steer him away from dwelling on what's done :-\
-
I can usually grasp at least a small part of things which are explained to me. I went dizzy trying understand what that was all about, I literally didn't have any clue. The problem is that some of this method of working out the statistical probability is subjective. You start with an estimate, a guess, and then try to prove it.
My lads aspergers brain doesn't work that way, it's a pure maths brain. When I asked him why he'd chosen that module, he said he thought it was statistics, which he loves. He obviously didn't research the Bayesian bit. There's a valuable lesson in that, I think.
-
Bayesian statistics is a system for describing epistemological uncertainty using the mathematical language of probability. In the 'Bayesian paradigm,' degrees of belief in states of nature are specified; these are non-negative, and the total belief in all states of nature is fixed to be one. Bayesian statistical methods start with existing 'prior' beliefs, and update these using data to give 'posterior' beliefs, which may be used as the basis for inferential decisions.
They lost me at 'describing'
And me..........!! :o :o :o
Not that I have any skills in advanced mathematics! ;D ;D ;D
Yes, a very BIG WELL DONE to STEMO's lad even studying that, let alone taking an exam on it at uni 8) 8) 8) :-* :) :)
-
Taking that is an achievement in its own right, and if he has any grasp of the concepts, then I am sure he will be pleasantly surprised with the result.
Regardless of how someone's mind works he cannot alter the result after the event and should already be looking forward to the next thing. Easier said than done, even for simple people like us, so I don't relish you having to steer him away from dwelling on what's done :-\
He's pretty good at accepting a situation, Al. Being philosophical is unusual for a lad his age. His saving grace, in my opinion, will be that the results are moderated. So, regardless of his actual mark, as long as he is 'in a range', he will pass. Add to that the fact that he has regularly achieved 90%+ in most of his other modules, I doubt he will be sanctioned in any way for this.
Most of his peers will graduate this year, he is hoping to do a masters next year and then a PhD. So I can understand his anxiety, but I'm pretty sure he'll be ok.
After about six weeks of fretting ;D
-
Obviously takes after his Mum, so he should do pretty well.
-
Obviously takes after his Mum, so he should do pretty well.
He's got my handsome good looks though ;D
-
Poor kid.
-
Brains from his mother, obviously.
No scouser is going to understand that. :)
-
Brains from his mother, obviously.
No scouser is going to understand that. :)
bet the lad can buzz the wheels off a car faster than the McLaren pit crew though :y
-
Brains from his mother, obviously.
No scouser is going to understand that. :)
bet the lad can buzz the wheels off a car faster than the McLaren pit crew though :y
it's a well known fact that all scousers are born with glue on their fingers. :)
-
I can see this being a really useful tool for predicting the 2019-nCoV infection & death rates from previous SAR & influenza pandemics prediction curve statistics where this is very, very difficult to work out until afterwards, due to it being an asymmetric statistics problem which is described well here:
https://towardsdatascience.com/why-everyone-knows-and-acts-like-the-2019-ncov-statistics-are-misleading-5919b3c33476 (https://towardsdatascience.com/why-everyone-knows-and-acts-like-the-2019-ncov-statistics-are-misleading-5919b3c33476)
Being able to use past known predictive distribution curves & what is known so far with this current pandemic you could produce a posterior distribution curve using Bayesian statistics & then as this pandemic develops see how the real world results compare to the calculations for an unknown confidence value in how accurate the results are.
That would be a good question for you to ask your lad STEMO.
-
Bayesian statistics is a system for describing epistemological uncertainty using the mathematical language of probability. In the 'Bayesian paradigm,' degrees of belief in states of nature are specified; these are non-negative, and the total belief in all states of nature is fixed to be one. Bayesian statistical methods start with existing 'prior' beliefs, and update these using data to give 'posterior' beliefs, which may be used as the basis for inferential decisions.
They lost me at 'describing'
Me too. ;D
-
I can see this being a really useful tool for predicting the 2019-nCoV infection & death rates from previous SAR & influenza pandemics prediction curve statistics where this is very, very difficult to work out until afterwards, due to it being an asymmetric statistics problem which is described well here:
https://towardsdatascience.com/why-everyone-knows-and-acts-like-the-2019-ncov-statistics-are-misleading-5919b3c33476 (https://towardsdatascience.com/why-everyone-knows-and-acts-like-the-2019-ncov-statistics-are-misleading-5919b3c33476)
Being able to use past known predictive distribution curves & what is known so far with this current pandemic you could produce a posterior distribution curve using Bayesian statistics & then as this pandemic develops see how the real world results compare to the calculations for an unknown confidence value in how accurate the results are.
That would be a good question for you to ask your lad STEMO.
I don't think he'd want to hear it ;D