Curve fitting, approximation, extra- or inter- polation, volatility, сonfidence interval

greenskull · 2021-10-06, 19:15

Hi, guys,

If you have a data set and you need to investigate it (see Title/Subj), then please don't hesitate to contact me.
Perhaps I can suggest a better solution here.
The quality of the fitting will be even better if you briefly comment on the nature of this dataset.

And I will do it for you with great pleasure :)

Batalov · 2021-10-06, 19:38

Go on then. Show them how it is done.

greenskull · 2021-10-06, 19:46

Quote:

Originally Posted by Batalov

Go on then. Show them how it is done.

I've already done that :)
https://mersenneforum.org/showthread...071#post583071

If your question is how I guessed that it was such a model, then the shape of the curve reminded me of the rank distribution curve. I investigated these curves some time ago. I just guessed and experimented with a data set and the model a little.

Batalov · 2021-10-06, 20:04

Where are the confidence intervals? Where is truth in advertising?

Don't bother answering though. These were rhetorical questions.

I have no further questions to anyone who is not aware of the appropriate precision of presenting the data.
Every time I see 15 digits of precision in slides, in presentation, aw well even in postings, - "I reach for my mental revolver." It shows immediately without any fear of contradiction the Dunning-Kruger effect in action.

greenskull · 2021-10-06, 20:22

Quote:

Originally Posted by Batalov

Where are the confidence intervals? Where is truth in advertising?

Don't bother answering though. These were rhetorical questions.

I have no further questions to anyone who is not aware of the appropriate precision of presenting the data.
Every time I see 15 digits of precision in slides, in presentation, aw well even in postings, - "I reach for my mental revolver." It shows immediately without any fear of contradiction the Dunning-Kruger effect in action.

I didn't count the confidence interval. The question was not posed that way. The dude asked to advise him the model. He was advised.
And in support of my advise, I showed the correlation coefficients.
You don't need to look at all the numbers, just look at five nines after zero and comma. It is enough.

2021-10-07, 01:57

https://www.curveexpert.net/
Been using it for years and there have been many others. You can "roll your own" after a while.
CERN had some interesting publicly available software that I was using in the mid 90's and I believe they still do, but I could be wrong.
Analyzing particle physics data and reporting on the same is a professional grade art.

greenskull · 2021-10-07, 08:56

Yes, it's a useful tool.
But this software does not perform spectral analysis. And it does not allow, for example, to identify harmonic components within the process.
And also it will not be possible to investigate complex Transient response. Much will have to be done manually.

Dr Sardonicus · 2021-10-07, 23:19

Quote:

Originally Posted by greenskull

Quote:

Originally Posted by Batalov

Go on then. Show them how it is done.

I've already done that :)
https://mersenneforum.org/showthread...071#post583071

I note that this "advice" came four months after the OP had stated

Quote:

Table-lookup param-selection is coded up.
<snip>
Thanks to all for playing!

greenskull · 2021-10-07, 23:49

Quote:

Originally Posted by Dr Sardonicus

I note that this "advice" came four months after the OP had stated

Alas, yes.
I really enjoy solving these problems. Therefore, I brought it out as a separate post here.
Maybe I can help someone and will do it in time.

2021-10-06, 19:15	#1
greenskull Xebeche Apr 2019 🌺🙏🌺 5·89 Posts	Curve fitting, approximation, extra- or inter- polation, volatility, сonfidence interval Hi, guys, If you have a data set and you need to investigate it (see Title/Subj), then please don't hesitate to contact me. Perhaps I can suggest a better solution here. The quality of the fitting will be even better if you briefly comment on the nature of this dataset. And I will do it for you with great pleasure :)

2021-10-07, 01:57	#6
jwaltos 7,649 Posts	https://www.curveexpert.net/ Been using it for years and there have been many others. You can "roll your own" after a while. CERN had some interesting publicly available software that I was using in the mid 90's and I believe they still do, but I could be wrong. Analyzing particle physics data and reporting on the same is a professional grade art. Last fiddled with by jwaltos on 2021-10-07 at 02:02

2021-10-07, 08:56	#7
greenskull Xebeche Apr 2019 🌺🙏🌺 5×89 Posts	Yes, it's a useful tool. But this software does not perform spectral analysis. And it does not allow, for example, to identify harmonic components within the process. And also it will not be possible to investigate complex Transient response. Much will have to be done manually. Last fiddled with by greenskull on 2021-10-07 at 09:25

Similar Threads
Thread	Thread Starter	Forum	Replies	Last Post
Odd ms/inter and ETA info on one machine	irowiki	Software	1	2018-09-25 13:38
Easier pi(x) approximation	mathPuzzles	Math	8	2017-05-04 10:58
Inter-Search Time	davar55	Forum Feedback	9	2012-12-09 15:41
Fitting rectangle	ccorn	Puzzles	27	2010-08-19 15:17
approximation of possible self-avoiding walks	ixfd64	Math	1	2007-11-26 05:57

2021-10-06, 19:38	#2
Batalov "Serge" Mar 2008 Phi(4,2^7658614+1)/2 23507₈ Posts	Go on then. Show them how it is done.

2021-10-06, 20:04	#4
Batalov "Serge" Mar 2008 Phi(4,2^7658614+1)/2 5×2,011 Posts	Where are the confidence intervals? Where is truth in advertising? Don't bother answering though. These were rhetorical questions. I have no further questions to anyone who is not aware of the appropriate precision of presenting the data. Every time I see 15 digits of precision in slides, in presentation, aw well even in postings, - "I reach for my mental revolver." It shows immediately without any fear of contradiction the Dunning-Kruger effect in action.