Disputatio:Distributio normalis

De valore expectato

Concerning the term "expected value" used by many statisticians, here rendered as valor expectatus, I think it is absurd term. Given many physically-relavent bimodal distributions the mean value of the distribution may observed with zero probability. Thus we would have the absurd situation where the "expected value" is literally the least expected value (observed with zero probability). In physics we use the term expectation value instead, which avoids somewhat this problem. I realize this isn't a latin question per say but I just wanted to make a point about it. It seems silly to continue this copy this illogical terminology into latin as well.--Rafaelgarcia 17:27, 22 Februarii 2008 (UTC)[reply]

Thank your for corrections: I said: See the disputatio page after having waited a minute. However, I have changed my mind. It is not necessary that everyone can find this discussion. Let's stay here:

Distribution function, which I called functio distributiva is a terminus technicus and should not be changed to a function that gives something. It change this back.

theorema, -tis is neutrum, so centralis relates to limitis. I thought it is related to theorema (from German: zentraler Grenzwertsatz, Grenzwert=limit, Satz = Theorem). English: central limit theorem, here central seems indeed to be relatet to limits, and I see that this is more meaningful, so the correction sounds well, but must also be used in the text, where limitivum has not yet been changed. I shall do this. Again: How should bell curve be translated? --Bachmai 17:26, 22 Februarii 2008 (UTC)

My best try at bell curve is curva campana forma--Rafaelgarcia 18:09, 22 Februarii 2008 (UTC)[reply]

That's definitely intelligible and might well be usable—though I just now find in Cassell's that campan- is non-Classical, and apparently the only term the Golden Age knew for 'bell' was tintinnabulum. IacobusAmor 18:44, 22 Februarii 2008 (UTC)[reply]

THe theorem is not central, but rather the theorem concerns a central limit, thus theorema de limite centali or theorema limitis centralis which is more usual in latin. Also the technical terms like "density function" in latin are more usual to render as densitatis functio, like in the Romance languages I believe, than to invent an adjective densitiva, or do you have an attestation for this term in latin? If one writes in latin, one has to try to think in latin and follow the norms of the language, otherwise we are just making things up.--Rafaelgarcia 17:51, 22 Februarii 2008 (UTC)[reply]

I changed densitatis functio ->probabilitatis densitas, which is more descriptive and closer to french. Please revert if you have a latin source.--Rafaelgarcia 18:28, 22 Februarii 2008 (UTC)[reply]

As you have seen, I have also done some changes. I think I have nothing overwritten except for what was intended by myself. It is much better to summarize the two sections as now since functio distributiva was in two sections before. Yor intention was to avoid the two distrib.. at "functio distributiva distributio normalis". This has completely succeeded now. "usus ea" = the use of it, i.e. the distributio canonica. "Numerus fortuitus" is not possilbe. It is the result X(omega) of variabilis fortuita. I hope that "sic distributa" is o.k. "ex distributione" is not said for a random variable, only for a number, which was not meant. Besides, I am not yet content with your formulation of the central limit theorem. The additional information is o.k. although I do not understand what elementum confertum means. My intention is: Not the normal distribution should be approximated, as it sounds now, but the true distribution by the normal.--Bachmai 19:39, 22 Februarii 2008 (UTC)[reply]

confertum = summarized. Feel free to fix as needed.--Rafaelgarcia 19:44, 22 Februarii 2008 (UTC)[reply]

Confertum (from confercire) = 'closely compressed, dense, stuffed with, full of'. Are you sure you don't want conlatum (from conferre) = 'brought together, collected'? IacobusAmor 20:03, 22 Februarii 2008 (UTC)[reply]

Ah yes you're right!--Rafaelgarcia 20:34, 22 Februarii 2008 (UTC)[reply]

I've just read your expectation/expected discussion. I would also prefer expectation value, although I did not think of your arguments mentioned. I used expected value because it was easier to translate in Latin. If you find a better expression, which means expectation, replace it please.--Bachmai 19:53, 22 Februarii 2008 (UTC)[reply]

OK i'll try. I also tried playing with the central limit theorem. Perhaps you can have a look and see if I missed the idea.--Rafaelgarcia 19:56, 22 Februarii 2008 (UTC)[reply]

One possibility is to render the idea as valor medius expectatus or valor medius theoreticus--Rafaelgarcia 19:58, 22 Februarii 2008 (UTC)[reply]

I find "de rebus" unsuited. Sums and means are what converges. You have omitted the identical distribution. This is not a necessary condition, but sufficient. And an exact mathematical formulation (as I had, this could be done in the article TLC (CLT) itself.) need not be intended. A more general version might even be better in this overviewing part of the CLT. But if you omit that all summed quantities have identical distribution, it does not make sense to talk about a single sigma square, because the variances of these quantities can be differnt. So it is perhaps better to only write something like "under some weak conditions"; such a sentence is always correct. The only thing that is violated in practice is the independence which urgently needs to be mentioned and is now mentioned. (I have seen that you have found my mistake about the inflection points; thanks.)--Bachmai 20:59, 22 Februarii 2008 (UTC)[reply]

I think "de rebus" is o.k. It is used in Latin for everything. To let the nice formulations I let it unchanged. I've corrected things about variances. It is now simpler. And I am content with the TLC as you have formulated it in essential. I changed the section Ad parametra aestimanda. "sufficiunt" was to weak for me. Mean and std.dev. are the only efficient ones. The grammar now used is very simple. However, the nice gerund construction is already in the headline, so there is no need to repeat it in the text. I still added the reason for the normal distribution: quia sub hac distributione valorem medium esse aestimatorem efficientem scivit. Is sub hac distributione o.k. "sub", "under", "unter" is in German and English o.k., but also in Latin? Otherwise change it please! --Bachmai 22:11, 22 Februarii 2008 (UTC)[reply]