User talk:Melcombe

Hello,

Thank you for the edit of my previous edit of the page on case-control studies. My apologies for not respecting the standard layout -I will try to do better in future!

As an epidemiologist however, I feel I must respectfully disagree with certain statements which are now part of this introduction. For instance, the statement that the "retrospective, non-randomized nature [of case-control studies] limits the conclusions that can be drawn from them" is, I feel, not entirely accurate. One aspect which in fact bothered me about the original version of this page was the suggestion that case-control studies are somehow necessarily inferior to cohort studies, as well as the somewhat irrelevant comparison with randomized controlled trials.

It would seem to me that there are two issues which are being confused here. The first being the fact that case-control studies are observational and thus do not constitute the same level of evidence as randomized trials. It does bear noting however that the same could be said of cohort studies. The second issue is the debate concerning the comparison between prospective cohort studies and case-control studies -i.e. which if any is 'better'? My discomfort here mainly concerns this second point: I have met several respected members of my field who do not feel that the case-control study is inferior to a cohort. In fact, depending on the situation and particular research question(s) of interest, the case-control study may even be better. In particular, case-control studies make it much easier to study certain conditions such as cancer, for which it takes a considerable amount of time for a sufficient number of events to accrue and for which cohort studies are frequently underpowered (and this despite many years of follow-up).

If anyone reads the talk pages (the last note dated on the case-control study one dated back from 2008 so I did wonder :) I would post my next edit there prior to integrating it into the actual page. (It just didn't seem to me that anyone was reading it which is why I posted my prior edit directly to the page itself.)Alic D (talk) 21:09, 3 August 2011 (UTC)[reply]

Dear Melcombe, I was wondering if you could explain why my modification to the Negative binomial mgf was wrong. I seem to have conflicting views from my university notes, textbooks and the internet. I have an exam tomorrow so would appreciate any explanation. Econstatgeek (talk) 18:05, 1 May 2008 (UTC)

Unfortunately I do not have time to do a full derivation. I have checked the formula in a couple of books. Also there must be a direct correspondence between the mgf and the characteristic function. ONe thing that confuses things is a tendency for different books to use different parameterisations of the negative binomial, so that the same correct formula can look different. Melcombe (talk) 08:42, 2 May 2008 (UTC)

I'll put it down to a difference in notation convention then. Thank you for your reply. Econstatgeek (talk) 16:11, 2 May 2008 (UTC)[reply]

Dear Melcombe,

I recently proposed starting a "probability and statistics" sub-project (aka task force or work group) of WikiProject Maths and was wondering if you'd be interested in participating. If so, please add your name and any comments at WP:WikiProject Council/Proposals#Probability and statistics. Regards, Qwfp (talk) 22:05, 12 March 2008 (UTC) (PS I believe it's traditional to start a user talk page with some variation on "Welcome to Wikipedia" but as you joined before I did I don't feel I'm in a position to welcome you. The lack of previous comments here would seem to reflect positively on your editing.)[reply]

I noticed that you put an "expert" tag on F-test and I completely agree with that. But I wonder if you can be specific about your concerns about order statistic. Michael Hardy (talk) 19:35, 10 April 2008 (UTC)[reply]

OK, I have placed something on article talk page. Melcombe (talk) 08:53, 11 April 2008 (UTC)[reply]

In my humble opinion, Statistical arbitrage is more within the scope of something like WP:WikiProject Finance than WP:WikiProject Statistics. It is a topic in Finance (or Mathematical Finance) having some statistical aspects, rather than a mainstream Statistics topic. Encyclops (talk) 17:58, 22 April 2008 (UTC)[reply]

See Wikipedia talk:WikiProject Statistics#Things on boundary of scope. My thought was that Statistical arbitrage did involve some statistical thinking in setting up the method. At present WikiProject Statistics is trying to bring all articles on statistics topics into the list of statistical topics to provide a basis for further progress. You might want to look at the project page and consider joining. Melcombe (talk) 18:09, 22 April 2008 (UTC)[reply]

Good idea for creating Category:Meta analysis. However, I do think that it should be spelled "meta-analysis". I have therefore requested a renaming to Category:Meta-analysis on WP:CFD. Please comment on this page. JFW | T@lk 13:12, 30 April 2008 (UTC)[reply]

Hello. Can you please explain the reason for the change in category? Thanks! -- Avi (talk) 14:38, 6 May 2008 (UTC)[reply]

Fair point. Thanks for taking the time to explain. -- Avi (talk) 17:20, 6 May 2008 (UTC)[reply]

Honorable Melcombe: Is it valid to estimate a confidence interval by multiplying the Student's critical T Value (e.g., +/- 1.96 for a 95% confidence range in a normal population) times the standard deviation of the data range? For example, if the Standard Deviation of the series happens to be 10 and the estimated value is 100, then the 95% confidence range would be from 81.4 (100 - 1.96 x 10) to 119.6 (100 + 1.96 x 10).

If this isn't the appropriate forum to ask this question, please forgive my ignorance and intrusion. Mike. --Grizthedog (talk) 21:50, 13 June 2008 (UTC)[reply]

You should really find a better forum for discussing all this. You might try the newgroup called sci.stat.math ... if you don't have a news-reader, it is accessible at http://groups.google.com/group/sci.stat.math/topics .

You need to be more clear about what it is you want to construct. The most usual case is a confidence interval for the mean, in which case your formula would be wrong as you would need to divide the standard deviation of the sample by the square root of the sample size. If you want to find a confidence interval for a further observation from the same population, then your formula in approximately correct ... for modest same sizes you would need to multiply the sample standard deviation by the square root of (1+1/N), where N is the sample size (you might find something about this in the article on "tolerance intervals" ). All this would be making the usual assumptions, but your use of the word "series" might indicate that you have a "time-series" in which there might be serial correlation, in which case the assumptions would not hold. Melcombe (talk) 09:07, 16 June 2008 (UTC)[reply]

I've replied to your post at WP:VPR.

The Transhumanist 19:58, 4 July 2008 (UTC)[reply]

As you have contributed to an earlier related discussion at Talk:St. Petersburg paradox#That d_mn'd period, you may be interested in Talk:St. Petersburg paradox#Request for comments: punctuation after displayed formula.  --Lambiam 18:17, 8 August 2008 (UTC)[reply]

Hello Melcombe. Your prod of K-factor error was contested. See Wikipedia:Articles for deletion/K-factor error. EdJohnston (talk) 00:44, 25 August 2008 (UTC)[reply]

The latest change is better (though the link doesn't work). However, the article still needs a simple definition of confidence level (i.e. something a layman can understand), and more importantly, an indication that, when used with statistical testing, it should be interpreted as indicating the significance level, preferably with a link to that article. The reason for this is simply the obvious fact that, for the average layman, the reason they would be looking up confidence levels or confidence intervals it that they have seen something in print reporting the results of statistics based research and want to know what the terms mean.

We can do the standard wikipedia "who can hold their breath longest" thing over this, but I'd rather be sensible about it. Perhaps if you changed the article to include these changes in the way you want, we can avoid the baby stuff.

Jim 125.255.16.233 (talk) 12:44, 12 October 2008 (UTC)[reply]

Replied on article Talk page. Melcombe (talk) 15:28, 14 October 2008 (UTC)[reply]

Please look at Talk:Central_limit_theorem#Asymptotic_normality_for_statistical_estimators. Boris Tsirelson (talk) 19:02, 5 November 2008 (UTC)[reply]

Hi Melcome. Thanks for your constructive edit to Chi squared distribution. It beats me why such a well-known fact has attracted such controversy. People (mostly but not exclusively anons) have repeatedly removed the assertion about normality. And it's not vandalism, as they justify their actions on the talk page. Do you have any insight in to why they do this? Me, I'm baffled. Best wishes, Robinh (talk) 20:46, 11 December 2008 (UTC)[reply]

Melcombe, wrt. the Pearson's_chi-squared_test definition, if you think notation outranks statistics, how can I argue with you? Muhali (talk) 02:38, 23 February 2012 (UTC)[reply]

It certainly does not help clarity if you use the same notation for two different things. Melcombe (talk) 02:58, 23 February 2012 (UTC)[reply]

True. So, is it better to use two notations or none?--Muhali (talk) 03:08, 23 February 2012 (UTC)[reply]

Note that I reverted most of your changes. Please take a look at the article talk page.--Muhali (talk) 18:12, 23 February 2012 (UTC)[reply]

Thank you, Melcombe, for your edit on the standard deviation estimate not being unbiased. I was concerned about this too and was contemplating a similar edit. However I still have some concerns and would value your opinion. Please see the student's t test talk page for details.

Best wishes, SciberDoc (talk) 22:20, 1 January 2009 (UTC)[reply]

Hi. I reverted a change to the exponential distribution article about the minimum of exponential variables, and then noticed that it was your change, and that you are a statistician, and I am not. This prompts me to notify you of my change, in case I am mistaken. I think that you ran afoul of the notation in the article, where the rate ${\displaystyle \lambda }$ is used in the pdf like this: ${\displaystyle \lambda e^{-\lambda x}}$, instead of the alternative definition of ${\displaystyle e^{-x/\lambda }/\lambda }$. -- Coffee2theorems (talk) 16:07, 8 January 2009 (UTC)[reply]

Hi Melcombe, you were the last person improving Cook's distance page. There is still one symbol that I am not sure what exactly it means: ${\displaystyle Y_{j(i)}}$. Could you please add its description to the page? Thanks, Alex -- talk 17:40, 17 March 2009 (UTC)[reply]

I have added something. Melcombe (talk) 17:52, 17 March 2009 (UTC)[reply]

Thanks for both the fix and your remarks to me.Alex -- talk 18:50, 23 March 2009 (UTC)[reply]

I see you've been listing lots of books as references, but in many cases without any specific citations (footnotes); please check out Wikipedia:Citing sources and ask me if you'd like help converting them to something more useful. Dicklyon (talk) 16:30, 27 March 2009 (UTC)[reply]

I've responded to your question on Talk:Arrival theorem and would be interested to hear you views on what you feel the most suitable title is. Whilst random observer property or job observer property both seem suitable to me, arrival theorem and PASTA property (for the Poisson process) seem to be more commonly used, judging by Google Scholar. Gareth Jones (talk) 11:44, 3 June 2009 (UTC)[reply]

Hi Melcombe, I wonder if you wouldn't mind if we carry on this discussion to your talk page? Anyways, this is what Lukacs has to say about the subject of mgf (i have his book right now, 2nd ed.):

On p.10 he defines an arbitrary integral transform of an r.v., and gives as an example kernels ${\displaystyle K_{1}(t,x)=e^{tx}}$ and ${\displaystyle K_{2}(t,x)=t^{x}}$ which give rise to the moment-generating function and probability-generating function respectively. These two functions are essentially the same, as ${\displaystyle {\text{mgf}}(t)={\text{pgf}}(e^{t})}$. He then remarks

Probability generating functions were introduced by Laplace; we will use these functions only rarely (in section 6.3) and mention them here mainly because they were the first integral transforms systematically used in probability theory.

On p.196 he returns to the subject of mgfs to mention that if f(z) is an analytic characteristic function then M(y) = f(-iy), where y is real. Note that Lukacs never defines mgf for complex or imaginary arguments.

On p.198 he gives an example of a function which has moments of all orders yet doesn't possess an mgf (or an analytic cf).

On p.251 he references a paper by Lévy(1937) where he used mgfs to establish certain convolution properties of Poisson-type distributions.

This is an exhaustive list. My interpretation of this list is that Lukacs doesn't really care much about mgfs, though it would have been a shame not to mention them if only but briefly.

And another question, Melcombe, when you say that "And considering that the major work on characteristic functions defines them for complex t it needs to be included in the main definition." -- could you please give a reference where you have seen such definition? Both Cuppens and Lukacs define it as

If p is probability measure on ${\displaystyle \mathbb {R} ^{n}}$, then ${\displaystyle {\hat {p}}}$ is analytic if there exists a function f defined on ${\displaystyle \mathbb {C} ^{n}}$ with complex values which is regular in some neighborhood of the origin and some constant δ such that

${\displaystyle {\hat {p}}(t)=f(t)\quad (t\in \mathbb {R} ^{n},\ \lVert t\rVert <\delta )}$

and later on they prove that if cf is analytic then it admits a representation

${\displaystyle {\hat {p}}(z)=\int _{\mathbb {R} ^{n}}e^{i\,(z,u)}p(du)}$ in a convex tube ${\displaystyle \mathbb {R} ^{n}+i\Gamma }$

My assessment of this theory is following: the characteristic function is defined only for real-valued arguments, because only then it has all the nice properties such as existence, and boundedness, and it is also much easier to generalize such cf to multi-dimensional rv's. It is also sometimes possible to build an analytic continuation of the cf to a horizontal strip ${\displaystyle \{-\alpha <\operatorname {Im} (z)<\beta \}}$, which allows us to derive additional results. However such continuation is only possible if r.v. has infinite number of moments and when the rate of growth of those moments is not too high.

Stpasha (talk) 22:19, 18 June 2009 (UTC)[reply]

OK, I have now had a little time to look through what sources I have access to, and they do take the approach of defining the cf for real arguments only, with a later extension to complex values by continuation using analytic functions, which they do say is equivalent to the basic Fourier-type integral for complex arguments (once it is known that the integral exists via the analytical continuation argument). This may be for historical reasons. There may be some way of developing the theory starting from the Fourier-type integral for complex arguments but I haven't found this set down. Nevertheless, because some of the most important results of cf theory are derived via this theory of extension into the complex domain, it seems important in an encyclopedia to note this at an early stage.

As for moment generating functions, it may be that they are unnecessary for theroeticians, the fact is that they are taught at a fairly elementary level of probaility and statistics whereas characteristic functions are not (not least because they involve imaginary numbers). And they are actually used with complex aruments in certain fields of application, where the mgf for some system is derived via theoretical arguments and converted to a density function using numerical inversion based on the inversion formula involving an integral parallel to the imaginary axis. Of course they could do all this using characteristic functions and an intergral along or parallel to the real axis, but they don't. I believe the shift in the axis of integration can help the numerical properties of the integration scheme.

Melcombe (talk) 10:09, 25 June 2009 (UTC)[reply]

I saw this only now. MGF and Laplace transforms have advantages for exponential families and especially for maximum-likelihood estimation from such families~(Barndorff Nielsen). More broadly, Laplace transforms preserve positivity and analyticity, while Fourier tranforms and characteristic functions don't.  Kiefer.Wolfowitz  (Discussion) 00:28, 22 February 2011 (UTC)[reply]

I trust that the introduction is improved. Do you have further feedback? ThanksKiefer.Wolfowitz (talk) 13:45, 29 June 2009 (UTC)[reply]

The split into two different pages I still think is a mistake, so I have raised this at WT:WPSTAT to try to get wider input. Jheald (talk) 17:27, 30 June 2009 (UTC)[reply]

I have nominated the article for deletion on the grounds stated. Please provide a prominent source per Wikipedia:Verification. I understand the motivation for having an article like this but question it on the grounds of this policy. I'll leave the content in regression to the mean for the moment but the linked article needs to sources for the content to remain. I am assuming good faith, I hope my concerns are clear. Please discuss if not. Cheers. Holon (talk) 12:58, 30 July 2009 (UTC)[reply]

The Cambridge Dictionary includes this: "The term is now generally used to label the phenomena that a variable that is extreme on its first measurement will tend to be closer to the centre of the distribution for a later measurement.

Does it really say "phenomena" instead of "phenomenon"? Michael Hardy (talk) 21:41, 1 August 2009 (UTC)[reply]

I noticed you are interested in statistics, and access to statistics articles.

Please take a look at these:

• Outline of statistics
  • Outline of probability
  • Outline of regression analysis

What is your initial impression?

What's missing?

Are they structured well? That is, do they present their respective subjects well?

How can we make them more useful?

For more information, see WP:OOK, WP:OUTLINE, WP:WPOOK.

For examples of well developed outlines on other subjects see Outline of Japan, Outline of Buddhism, Outline of robotics, and Outline of cell biology, and Outline of forestry.

I look forward to your reply.

The Transhumanist 00:32, 16 September 2009 (UTC)[reply]

Your revert of my edits seem to be a bit hasty, not only are you incorrect in your summary later edits incorrectly changed meaning & number was correct as prob was for either 3 heads or three tails which clearly none of my revisions altered (see my diffs), but you have reverted my correction for the probability of 3 heads or tails which is 0.5*0.5*0.5 = 1/8 or an eighth not a quarter. --Zven (talk) 08:46, 24 September 2009 (UTC)[reply]

Right I now see what you are getting at, however you also reverted my other edits which probably didn't need to go. --Zven (talk) 08:57, 24 September 2009 (UTC)[reply]

Actually, the example is misleading, no wonder there is so much confusion, it states a two headed coin, so three tails is not an outcome at all, only three heads. --Zven (talk) 09:01, 24 September 2009 (UTC)[reply]

Doh! I knew that! :) Mia Culpa. Cheers. Jwesley78 (talk) 14:25, 22 October 2009 (UTC)[reply]

You're right. Thanks! Quantling (talk) 17:32, 17 November 2009 (UTC)[reply]

Should we have a separate article for the notion of total variation in probability theory and measure theory? I don't like the current arrangement where the material on probability theory has been forked off from the material on measure theory, even though they are essentially the same thing. What seems better to me is to have an article total variation in measure theory and probability theory that deals with both of these notions. Sławomir Biały (talk) 12:15, 9 December 2009 (UTC)[reply]

A general technique that is used in statistical methods applied to geochronology.

The mathematics provided deal with general cases, split into the two typical ways MSWD is utlized.

First, to provide a way to analyse the outcome of repeated attempts to measure the same "age", each attempt associated with a known variance (obtained by repeated sampling during the individual attempt. This assumes the sample has a single age.

Second, to provide a way to analyse an isochron, usually defined on a plot of two ratios (e.g. 36Ar/39Ar versus 39Ar/40Ar). —Preceding unsigned comment added by 202.55.153.53 (talk) 09:10, 7 January 2010 (UTC)[reply]

That was a nice edit of the contingency table page. Thanks. Iss246 (talk) 22:36, 11 January 2010 (UTC)[reply]

Please, before reverting the edit and asserting that I “completely ignored the discussion on the talk page”, make sure you have read that discussion. There was a 1-month old proposal (section Major Changes) to simplify the exposition down to only 1 main definition, and that proposal was met with a (cautious) support.

The main reason why we had 2-column infobox is because readers were frequently confused when they seen apparent discrepancies between the article and their textbooks. Such confusion can be avoided either by bloating the infobox (and actually there are more than 2 possible definitions), or by including a very noticeable alertbox, warning about potential discrepancies when comparing this info to existing textbooks, which is the way we are dealing with the problem right now.  … stpasha »  18:55, 19 January 2010 (UTC)[reply]

See User talk:Stpasha for reply. Melcombe (talk) 10:28, 20 January 2010 (UTC)[reply]

Hi Melcombe, please see talk:Sigma-algebra#Too technical?, thanks. Paul August ☎ 15:18, 20 January 2010 (UTC)[reply]

Just my 2p: the {{context}} template is a lot more self-explanatory (to me at least) for an article where the lede needs to be expanded to explain what the topic is about. — Carl (CBM · talk) 20:20, 20 January 2010 (UTC)[reply]

Somehow this mangled sentence found its way into the article:

The consequence of "Studentization" is tcentralhat the complication of treating the probability distribution of the mean,

Can you clarify? Michael Hardy (talk) 23:10, 4 February 2010 (UTC)[reply]

This is an automated message from CorenSearchBot. I have performed a search with the contents of Pietro Balesta (economist), and it appears to be very similar to another Wikipedia page: Pietro Balestra. It is possible that you have accidentally duplicated contents, or made an error while creating the page— you might want to look at the pages and see if that is the case. If you are intentionally trying to rename an article, please see Help:Moving a page for instructions on how to do this without copying and pasting. If you are trying to move or copy content from one article to a different one, please see Wikipedia:Copying within Wikipedia and be sure you have acknowledged the duplication of material in an edit summary to preserve attribution history.

This message was placed automatically, and it is possible that the bot found similarity where none actually exists. If that is the case, you can remove the tag from the article and it would be appreciated if you could drop a note on the maintainer's talk page. CorenSearchBot (talk) 16:32, 9 February 2010 (UTC)[reply]

Thank you for your contributions to the encyclopedia! In case you are not already aware, an article to which you have recently contributed, Climate change, is on article probation. A detailed description of the terms of article probation may be found at Wikipedia:General sanctions/Climate change probation. Also note that the terms of some article probations extend to related articles and their associated talk pages.

The above is a templated message. Please accept it as a routine friendly notice, not as a claim that there is any problem with your edits. Thank you. -- TS 17:40, 9 February 2010 (UTC)[reply]

Thanks for correcting my spelling, which was surprisingly bad. Thanks also for expanding the bibiographical details of Freedman's Statistics. Best regards, Kiefer.Wolfowitz (talk) 14:33, 12 February 2010 (UTC)[reply]

Melcombe, you purged the category of procedures whose text already named them as obsolete. I don't understand your comment that categories shouldn't have just one item, given your purging behavior of Box-Pierce (whose text described it as obsolete) and the F-test. Kiefer.Wolfowitz (talk) 20:13, 8 March 2010 (UTC)[reply]

Hi there. Thanks for the post. It seems that everything here is done by editing a page. I hope this is the right way to "reply" a message. I didn't see any post on the discussion page about LAD, only those links to the stats groups, so there I went! Thanks for the pointers, I will write this down later.

Best,

--Gpfreitas (talk) 13:13, 9 March 2010 (UTC)[reply]

Hi, I saw that you edit stats pages. Is there an ambulance/ER type list for articles in real trouble, e.g. Random sequence is in bad shape, Statistical randomness is barely ok, but has errors, and both are almost reference free. Applications of randomness probably wins the worst article award and Randomness is no gem either. Is there a "help needed" list somewhere? Thanks. History2007 (talk) 13:30, 10 March 2010 (UTC)[reply]

Thanks, I think I will just add the templates on the pages. These pages are not my priority, I was just amazed how low quality they all are - compared with info on Paris Hilton that gets great attention in Wikipedia. Thanks. History2007 (talk) 16:35, 10 March 2010 (UTC)[reply]

Dear Melcombe,

I was wondering if you have a reference for the derivation of the probability of locating the ${\displaystyle N^{th}}$ nearest neighbor at some radial distance r, shown at the Complete spatial randomness page? A more detail outline of the derivation would be greatly appreciated.

Thank you very much,

Bsesar (talk) 21:46, 8 April 2010 (UTC)[reply]

Dear Melcombe, I trust that you are in good health and spirits. I have been a fan of your editing for a long time. I am sorry for my part in sometimes contentious editing on "statistical inference", but I think that the resulting article was improved, certainly over my rather crude initial contributions. Sincerely, Kiefer.Wolfowitz (talk) 22:36, 14 April 2010 (UTC)[reply]

Dear Melcombe,

The Koay inversion technique is indeed a parameter estimation if viewed from a statistical point of view but it is also an inversion technique when viewed from the "inverse problems" point of view. Of course, there are other estimation techniques but as an inversion technique the Koay inversion technique is very original and peculiar at the same time. As far as I can tell, I have not seen anything like this before. Your comment on the technique would be a welcome addition to the Wikipedia. Therefore, I think a more specific name for this inversion technique is not at all inappropriate.

-Carl —Preceding unsigned comment added by StanfordCommSci (talk • contribs) 16:52, 14 May 2010 (UTC)[reply]

Hi Melcombe! I updated the references and modified some statements in the article statistics education. Perhaps you could add something to the stub Making Mathematics Count, about the report of Adrian Smith (academic). Thanks, Kiefer.Wolfowitz (talk) 13:12, 15 May 2010 (UTC)[reply]

Hello, I'm still not sure why you think a composite null hypothesis is ok. You can't compute the probability of a type I error if your null hypothesis is composite. I have gone into greater detail on the Statistical Hypothesis Testing talk page.

Tank (talk) 09:55, 19 May 2010 (UTC)[reply]

Hello Melcombe! I appreciate your retaining the mean-unbiased and median-unbiased estimators in your latest edits (which only partially reverted mine). Could you (briefly) explain why you keep removing "questionaire" and "survey" from social statistics? (IMHO, surveys are intrinsically human-subjects statistics.) Also, I am sincerly puzzled by your repeated comment about the equality of sampling and experiments: Please explain a bit! Thanks, and best regards, Kiefer.Wolfowitz (talk) 19:16, 25 May 2010 (UTC)[reply]

Hi Melcombe - Thanks for questioning the validity of references at that article. Actually, the 2003 article does not use the term, either. It's easy to check because it is a free journal. Apparently, the term is introduced by the 2010 reference. And not only is that not independent, it is not verifiable, because the Publisher's website says it won't be released until August 30. Nor is it peer-reviewed. But I'm not sure whether this is a case for AfD or speedy, that's up to you.

Also: watch out for possible sockpuppets, like 76.112.241.229. Best regards, Iulus Ascanius (talk) 19:47, 1 June 2010 (UTC)[reply]

I have added the "reviewers" property to your user account. This property is related to the Pending changes system that is currently being tried. This system loosens page protection by allowing anonymous users to make "pending" changes which don't become "live" until they're "reviewed". However, logged-in users always see the very latest version of each page with no delay. A good explanation of the system is given in this image. The system is only being used for pages that would otherwise be protected from editing.

If there are "pending" (unreviewed) edits for a page, they will be apparent in a page's history screen; you do not have to go looking for them. There is, however, a list of all articles with changes awaiting review at Special:OldReviewedPages. Because there are so few pages in the trial so far, the latter list is almost always empty. The list of all pages in the pending review system is at Special:StablePages.

To use the system, you can simply edit the page as you normally would, but you should also mark the latest revision as "reviewed" if you have looked at it to ensure it isn't problematic. Edits should generally be accepted if you wouldn't undo them in normal editing: they don't have obvious vandalism, personal attacks, etc. If an edit is problematic, you can fix it by editing or undoing it, just like normal. You are permitted to mark your own changes as reviewed.

The "reviewers" property does not obligate you to do any additional work, and if you like you can simply ignore it. The expectation is that many users will have this property, so that they can review pending revisions in the course of normal editing. However, if you explicitly want to decline the "reviewer" property, you may ask any administrator to remove it for you at any time. — Carl (CBM · talk) 12:33, 18 June 2010 (UTC) — Carl (CBM · talk) 12:49, 18 June 2010 (UTC)[reply]

Melcombe,

You deleted my link to modecube.com saying that the link was of no use -which is rather a harsh statment-.

Please note that the first editor, Debivort, said it wasn't a clear call.

Then, people who don't have a PhD in statistics may have some legit questions such as where the covariance formula comes from, why the variance is derived from the covariance etc.

So, if:

• those question are indeed questions some people might have,
• there are no simple answers to that in the article and the existing links,
• the explanations in modecube.com make sense,

then I dont see why the link is of no use.

I may be wrong, I just would like to understand why (my goal is just to help as many people as I think those explanations would have helped me so much when I was a student)

Regards, Philippe 86.183.200.206 (talk) 13:33, 23 June 2010 (UTC)[reply]

Dear Melcombe,

Thanks you for your welcolme, your useful links and your answer.

My understanding of the later is that my points are valid (questions one can ask, answers not in article, answers in the external website) but that the explanations look like textbook, they are "how to" (calculate) material, and thus should not belong to Wikipedia but rather to some Wiki etc.

I can understand that, even if I would advocate that Wikipedia could give some further explanations/example when a subject is complex. For example, an article that helped me a lot when studying for the CFA exam is http://en.wikipedia.org/wiki/Bond_duration. And this article displays a short proof/explanation for the modififed duration.

What's more, I do believe that, even if Wikipedia should keep to an introductional level, it still could/should offer some further reading leads. As a case in point, programming articles usually provides in depth examples and links to detailed implementations (http://en.wikipedia.org/wiki/Factory_method_pattern etc.)

Best Regards, Philippe Vergnetp (talk) 08:40, 24 June 2010 (UTC)[reply]

Anova treatments are typically nominal. A Bernoulli response is often nominal. It is fallacious to say that experiments are quantitative, although this fallacy has been common in the lower regions of some social-science disciplines, particularly under the influence of 19th Century German fallacies. Please reconsider your adding "quantitative research" as a category for experiments.

It would be appropriate to put experiments under "mathematical methods" or "statistical methods", which are wider categories than "quantitative methods". Bets regards , Kiefer.Wolfowitz (talk) 15:06, 10 July 2010 (UTC)[reply]

Respected Melcombe,

I have looked at the description given for V-Optimal Histograms, Could you kindly elaborate how the weighted variance is calculated in the examples. It's formula is given as W = sum(nj, Vj). But using this formula I am not able to get the weighted variance of 2.28 for the given data. Please clarify it as soon as possible.

Thanks, Nagesh Bhattu,PhD Student —Preceding unsigned comment added by 220.225.248.2 (talk) 13:17, 15 July 2010 (UTC)[reply]

I don't know. I merely copied a formula from one of the papers that were cited but which no longer seem to be available online. There may a problem with know which version of variance to use (divided by N or N-1). You might try contacting the oroginal author of the article. Melcombe (talk) 15:29, 15 July 2010 (UTC)[reply]

This is an automated message from CorenSearchBot. I have performed a web search with the contents of John C. Gittins, and it appears to include a substantial copy of http://www.stats.ox.ac.uk/people/academic_staff/john_gittins. For legal reasons, we cannot accept copyrighted text or images borrowed from other web sites or printed material; such additions will be deleted. You may use external websites as a source of information, but not as a source of sentences. See our copyright policy for further details. (If you own the copyright to the previously published content and wish to donate it, see Wikipedia:Donating copyrighted materials for the procedure.)

This message was placed automatically, and it is possible that the bot is confused and found similarity where none actually exists. If that is the case, you can remove the tag from the article and it would be appreciated if you could drop a note on the maintainer's talk page. CorenSearchBot (talk) 17:10, 15 July 2010 (UTC)[reply]

Hi, why did you revert my changes? What is the definition of E_B and Pr(A|B) where B is a random variable? B must be an event, or in general a sigma-algebra, not a random variable. —Preceding unsigned comment added by 121.102.42.138 (talk) 13:29, 10 September 2010 (UTC)[reply]

You said ... You said ... Hi, why did you revert my changes? What is the definition of E_B and Pr(A|B) where B is a random variable? B must be an event, or in general a sigma-algebra, not a random variable. —Preceding unsigned comment added by 121.102.42.138 (talk) 13:29, 10 September 2010 (UTC)

       The line immediately above defined B as a random variable. I will change it back to X to avoid more confusion. But introducing =n here was wrong because n has a different meaning later, and the quantity condioned on has to be a random variable, otherwise the expectation can'y be taken. Melcombe (talk) 13:53, 10 September 2010 (UTC) 

Uhm... it is also true that Pr(A|X) is not defined if X is a random variable. It needs to be an event or (more generally) a sigma-algebra - in this case a sigma algebra generated by X. —Preceding unsigned comment added by 121.102.42.138 (talk) 14:31, 10 September 2010 (UTC)[reply]

Hello Melcombe. As a newbie, I'd like to thank you for helping out in cleaning up the Multifactor Design of Experiments Software article. If you have more suggestions, I'd be most welcome to follow your advice. Again, thanks! Agatecat2700 09:35, 13 September 2010 (UTC) —Preceding unsigned comment added by Agatecat2700 (talk • contribs)

Thank you for tidying up Category:Monte Carlo methods by creating Category:Monte Carlo methodologists. Fintor (talk) 10:48, 14 September 2010 (UTC)[reply]

The article E-net (metric spaces) has been proposed for deletion because of the following concern:

Doesn't seem to provide any more than a dictionary definition, albeit a technical one. WP:NOTDICT.

While all contributions to Wikipedia are appreciated, content or articles may be deleted for any of several reasons.

You may prevent the proposed deletion by removing the {{dated prod}} notice, but please explain why in your edit summary or on the article's talk page.

Please consider improving the article to address the issues raised. Removing {{dated prod}} will stop the proposed deletion process, but other deletion processes exist. The speedy deletion process can result in deletion without discussion, and articles for deletion allows discussion to reach consensus for deletion. GiftigerWunsch [TALK] 12:48, 13 September 2010 (UTC)[reply]

Could you please have a look at the changes Joe The-tenth-zdog made in the article statistical hypothesis testing. There are errors in them, inconsistencies, a strange way of introducing and superfluous tables. Nijdam (talk) 14:09, 23 September 2010 (UTC)[reply]

Thanks for interfering. Nijdam (talk) 22:27, 23 September 2010 (UTC)[reply]

Just thought I'd drop you a line that I undid your recent correction. I think the original time series formulation was correct and I added a reference supporting it. The original way also conforms better with the preceding equation. Do you want to add an additional formulation? --Bequw (talk) 05:07, 29 September 2010 (UTC)[reply]

Hello, how do you do? I also work in statistics (kind of) in the UK. Many thanks for the Significance page. I just overwrote a recent correction you made as I feared it could sound a little advertise-y..what do you think? Also, how could I change the title from 'journal' to 'magazine'? As you're the initial author of the piece would it be you who could change this or must I contact an editor? Thank you. Abdelhk (talk) 05 November 2010 15:11 (UTC)

I made a few edits to Nassim Nicholas Taleb and tagged it. I noticed your remarks on the closed AfD merge discussion and wonder if you'd be interested in working on the article to bring it up to GA or at least not a collection of fan club effusions. I've already gotten some push-back from one of the keepers of the flame, and it's something I woouldn't want to try to do alone. Tom Reedy (talk) 13:36, 9 December 2010 (UTC)[reply]

Hi Dr. Melcombe,

I recently have come across the "Poisson Binomial Distribution", which turns out to be something that greatly helps with a statistical problem I am currently dealing with. I know for a fact that a number of statistics textbooks do not have this distribution listed and Google does not really return hits that describe this distribution. When I showed the distribution to a few colleagues in the Math Department, they claimed they've never seen this distribution before as well (and they are pretty smart individuals). Since I am not a statistician myself, I don't exactly know how well-known this distribution is. Given that a paper about it was published in 1993 by Wang, I'd presume there would've been 17 years for statisticians to absorb this distribution into the fold. Do you think you can tell me how often this distribution is used in practice and whether or not it is accepted by the statistics community? If it is relatively obscure, then I might have to spend a bit of time to check over the math just to make sure I am not using something that's wrong (although on the surface, the pdf and mean equations look reasonable).

Thanks. Bobthefish2 (talk) 08:04, 20 December 2010 (UTC)[reply]

There may be some confusion here. The prime source I have uses "Poisson Binomial Distribution" for a different distribution than that set out in Poisson binomial distribution. However, it does have that distribution under the name "Poissonian binomial sampling", but does say that has been known also under that "Poisson Binomial Distribution" and as the "binomial distribution of Poisson". References go back to Poisson in 1837. My source is: Johnson NL; Kotz, S.; Kemp, A.W.(1993) Univariate Discrete Distributions, 2nd edition. Wiley ISBN 0-471-54897-9 (page138 for "Poissonian binomial sampling"). I think that perhaps this hasn't been generally counted as a "distribution" because it has so many parameters, but it has certainly been studied. Melcombe (talk) 10:00, 21 December 2010 (UTC) (Placed on user's talk page also) Melcombe (talk) 10:03, 21 December 2010 (UTC)[reply]

Dear Melcombe,

You have repeatedly removed the Mathematical Reviews template from (my) citations. Would you please cease removing this information, or argue against this template at the its talk page (with a notice at the Mathematics Project)?

Many readers of mathematical articles are at universities or institutes with subscriptions to Mathematical Reviews, and these readers benefit from the links. I know that I do! Even when a book or an article has been merely listed without a review, the Mathematical Reviews database provides helpful links to publisher, other articles by the author, forward citations to other references (sometimes later editions with reviews), and sometimes backwards citations, all of which are helpful.

(I have tried to improve my citation practice, following your suggestions. Please see my best efforts at the article on the Shapley-Folkman lemma, which is under good article review, and which would benefit from your scrutiny.)

Sincerely, Kiefer.Wolfowitz (talk) 11:49, 21 December 2010 (UTC)[reply]

Harvard citation templates[edit]

Dear Melcombe,

You might like the harvard citation templates that are used in the Shapley-Folkman lemma article. They allow page referencing with links to the original citation information.

Best regards, Kiefer.Wolfowitz (talk) 22:33, 19 January 2011 (UTC)[reply]

Thanks for tidying up DoE[edit]

Hi Melcombe,

I appreciate your updating the citations on DoE, particularly after you have expressed concern that I should consider whether Peirce needs to be cited so much (if my memory is correct). Your updating the Peirce citations is another sign of your good character!

Best regards,  Kiefer.Wolfowitz  (Discussion) 14:53, 15 March 2011 (UTC)[reply]

Hi, I wondered if you could explain your reasoning in changing categories for example in V-statistic. You changed Category:Statistics + Category:Probability + Category:Theory of probability distributions to just Category:Estimation theory.

I don't think that Estimation theory adequately covers the scope of V-statistis as they are not used just for estimation. In fact, in the many examples involving degenerate kernels typically the statistic is used for hypothesis testing and does not have a good application for estimation (e.g. Cramer-von Mises type statistics, and others). I could give more examples. Also I don't understand why general categories such as Statistics or Probability should be deleted. V-statistics encompass a wide class of statistics, so I thought category Statistics is applicable. But then I noticed that you called it a "silly" category for U-statistics. Perhaps I do not understand the rationale of a category. In biographies all categories that apply are used, even if one or more is a superset of the other. Why is it different here? When I pull up a page of categories on say Theory of probability distributions, I would hope that I can find most relevant articles and not just the ones that failed to fall into a more specific category. Have I missed the point? A good explanation would help me. Mathstat (talk) 20:20, 22 February 2011 (UTC)[reply]

Hi, thanks for your reply. Maybe the real problem is that there is not a broad enough yet narrow enough category for some articles. I replied more on my talk page below your explanation. Mathstat (talk) 17:48, 25 February 2011 (UTC)[reply]

Probabilities which are one another's reverse is meaningless, I think. Bayes theorem is about updating of beliefs given new evidence. Please explain what you think was wrong in my edit. Richard Gill (talk) 21:01, 3 March 2011 (UTC)[reply]

I have made another attempt at improvement. Richard Gill (talk) 09:33, 4 March 2011 (UTC)[reply]

Hi, Melcombe, thanks for redoing mean square error as mean squared error. Probably this has already occurred to you, but I hope you can do the same thing with Root mean square deviation. Duoduoduo (talk) 18:11, 16 March 2011 (UTC)[reply]

Hi, I wikified the whole article, and considered the fact that this is about a respectable Dutch professor. I don't think there is a need to restore those tags unless the original editor starts making mayor changes again. -- Mdd (talk) 19:14, 22 March 2011 (UTC)[reply]

please check Talk:Completeness_(statistics). I mean would you please provide an example that a complete statistic is not sufficient. We know sufficiency does not imply completeness, but can you provide an example that a minimal sufficient statistic which is not complete. So in total 2 examples. Jackzhp (talk) 15:06, 25 March 2011 (UTC)[reply]

Hi Melcombe. As you are one of the most active wiki-mathematicians and you once made some edits to the article Copula (statistics), I would like to ask you about something. I would like to develop and improve the latter article. Started with that on sunday, but got completely reverted by some guy who calls probability theory a small subset of statistics... I dont want to start an edit war. could you have a look at my edits and comment? or even better, if you have some time free, help editing? that would be great! I dont insist on the renaming, but the content revertions i do not understand. regards, Philtime (talk) 19:05, 17 May 2011 (UTC)[reply]

Hi melcombe, thank you very much for your constructive comments. I'll work over my edits in the near future according to the stated manual. Philtime (talk) 16:41, 21 May 2011 (UTC)[reply]

An article you deleted following an AfD has been restored (not by me). Is it possible that you could restore the old version of the Talk page. The detail for the Talk page seems to be:

• 12:56, 11 June 2009 Stifle (talk | contribs) deleted "Talk:Lukaszyk-Karmowski metric" ‎ (Wikipedia:Articles for deletion/Lukaszyk-Karmowski metric)

This is probably not very important but it might be helpful to have the old discussion. Melcombe (talk) 16:03, 18 May 2011 (UTC)[reply]

Done, although in future you will find you will get requests like this completed more quickly at WP:REFUND. Stifle (talk) 09:13, 19 May 2011 (UTC)[reply]

Hi Melcombe. Thank you very much for your edits - most helpful. There are now at least 3 links pointing to SABRE Research UK. How does the page get deorphaned? SBGREEN. P.S. I am very new to Wiki and any help or advice most welcome. —Preceding undated comment added 15:07, 2 June 2011 (UTC). PPS. Sorry forgot tilde thing - will add now. SBGREEN (talk) 15:16, 2 June 2011 (UTC)[reply]

Alright I saw you reinsert tags and added a few new ones. So what exactly is the problem? I don't quite see we need inline references for an article that at best 1/3 of page unless there is something controversial in there. If that's the case however it should be pointed out on the discussion page.--Kmhkmh (talk) 17:29, 9 June 2011 (UTC)[reply]

It's best not omit the hyphens that cue the reader to the structure of a phrase, in an encyclopedia that's intended to be as clear as possible, so could you fix your new categories, in which the compound nouns time domain and frequency domain are used without hyphens as adjectives modifying analysis? Dicklyon (talk) 14:24, 16 June 2011 (UTC)[reply]

Dear Melcombe,

I just wanted to ask you whether you can explain to me what has happened to the Jacobian during the derivation of the dirichlet distribution from gamma distributed variables (in the proof you have removed in the article 'Dirichlet ditribution')? I think it's more or less the same proof as in Luc Devroyes book referenced in that article, but I don't understand this one either.

Much thanks in advance — Preceding unsigned comment added by Tuonawa (talk • contribs) 11:43, 17 June 2011 (UTC)[reply]

This is an automated message from CorenSearchBot. I have performed a web search with the contents of Convolution of probability distributions, and it appears to include material copied directly from http://www.territorioscuola.com/wikipedia/en.wikipedia.php?title=List_of_convolutions_of_probability_distributions.

It is possible that the bot is confused and found similarity where none actually exists. If that is the case, you can remove the tag from the article. The article will be reviewed to determine if there are any copyright issues.

If substantial content is duplicated and it is not public domain or available under a compatible license, it will be deleted. For legal reasons, we cannot accept copyrighted text or images borrowed from other web sites or printed material. You may use such publications as a source of information, but not as a source of sentences. See our copyright policy for further details. (If you own the copyright to the previously published content and wish to donate it, see Wikipedia:Donating copyrighted materials for the procedure.) CorenSearchBot (talk) 13:05, 29 June 2011 (UTC)[reply]

The article Confidence in statistical conclusions has been proposed for deletion because of the following concern:

No sources, OR, reads like an essay

While all contributions to Wikipedia are appreciated, content or articles may be deleted for any of several reasons.

You may prevent the proposed deletion by removing the {{proposed deletion/dated}} notice, but please explain why in your edit summary or on the article's talk page.

Please consider improving the article to address the issues raised. Removing {{proposed deletion/dated}} will stop the proposed deletion process, but other deletion processes exist. The speedy deletion process can result in deletion without discussion, and articles for deletion allows discussion to reach consensus for deletion. --Nuujinn (talk) 18:37, 4 July 2011 (UTC)[reply]

Melcolme, thanks for cleaning up the categories on Marketing science . (I'm a newbie when it comes to new Wikipedia articles). I inherited the categories from Operations research , from which Marketing science has a heritage, so I don't appreciate whether the categories (e.g. Applied mathematics) apply or not. Marketing science without mathematics (wouldn't be much of a science). Daviding (talk) 15:04, 5 July 2011 (UTC)[reply]

Melcolme, I disagree with your reversion of an edit of mine. The lead that you have reverted to is simply incorrect. Progressively measurable processes can "see into the future": case point: deterministic processes can be progressively measurable. I'm not aware of any intuitive definition of it. The most that can be said is that it is a property that certain stochastic processes have. RobertHannah89 (talk) 10:15, 5 August 2011 (UTC)[reply]

Right, but let's just be clear, you're reverting an edit because the FALSE information met the understandability criterion. I have asked lecturers and I have trawled the internet and I cannot find an intuitive definition of progressively measurable. It is better to have a correct but difficult definition than a completely false one. Would you please undo your reversion?RobertHannah89 (talk) 00:52, 11 August 2011 (UTC)[reply]

Please notice my recent edit to stochastic equicontinuity. \Pr is a standard operatorname, and on some browsers looks different from \text{Pr}. Michael Hardy (talk) 11:17, 17 August 2011 (UTC)[reply]

You wrote:

Let ${\displaystyle \{H_{n}(\theta ):n\geq 1\}}$ be a family of functions defined from ${\displaystyle \Theta \rightarrow \mathbb {R} }$, where ${\displaystyle \Theta }$ is any normed metric space. Here ${\displaystyle \{H_{n}(\theta )\}}$ might represent a sequence of estimators applied to datasets of size n given that the data arises from a population for which the parameter indexing the statistical model for the data is θ. Then ${\displaystyle \{H_{n}\}}$ is stochastically equicontinuous if, for every ${\displaystyle \epsilon >0}$, there is a ${\displaystyle \delta >0}$ such that:

${\displaystyle \lim _{n\rightarrow \infty }\Pr \left(\sup _{\theta \in \Theta }\sup _{\theta '\in B(\theta ,\delta )}|H_{n}(\theta ')-H_{n}(\theta )|>\epsilon \right)<\delta .}$

Here B(θ, δ) represents a ball in the parameter space, centred at θ and whose radius depends on δ.

What exactly do you mean by the term "normed metric space"? Do you mean a normed vector space? Or maybe just a metric space?

How can a family of functions from the parameter space to the reals "represent a sequence of estimators"? Generally an estimator depends on data and not on the parameter, since an estimator is observable and a parameter being estimated is not.

Later you write of the probability that H_n(θ) differs from H_n(θ') by more than a certain amount. That seems to presuppose that H_n(θ) does not simply depend on θ, but also depends on the data, so it has a random component. That being the case, it seems misleading to say earlier in the paragraph simply that H_n is a function from the parameter space into the reals. Michael Hardy (talk) 11:52, 17 August 2011 (UTC)[reply]

The text here was mainly copied from article equicontinuity, and the problems originated there. Melcombe (talk) 08:29, 18 August 2011 (UTC)[reply]

There are generalizations of absolutely homogeneous norms (B* norms or for complete spaces B norms) to Frechet spaces and more generally to metric linear spaces, where they are called F* norms (or F-norms for complete spaces). (More generally, there are modular spaces, etc.) Kibbitzingly,  Kiefer.Wolfowitz 16:49, 16 September 2011 (UTC)[reply]

Surprising how much more balanced that little-to-start-with article looks with the "Scope" heading. --Thomasmeeks (talk) 18:08, 20 August 2011 (UTC)[reply]

Hi Melcombe, A few days ago I noticed this entry on your talk page and decided to take a look at economic methodology. I found a page with great sources but no content. I decided to make it into a bibliography, which is what it really was. However, now I find myself in need of guidance as to how to jump start a real article for this topic, and reformat the old article as a bibliography. This is an important field, which overlaps with statistics quite a bit, but as most economists, I do not know much about it. Any suggestion would be appreciated. Yaniv256 (talk) 19:55, 28 July 2012 (UTC)[reply]

Please review the map, the work of Tofanelli et al, Hassan et al, and comment in the discussions:

http://en.wikipedia.org/wiki/File:HG_J1_(ADN-Y)
http://hpgl.stanford.edu/publications/AJHG_2004_v74_p1023-1034.pdf http://ychrom.invint.net/upload/iblock/94d/Hassan%202008%20Y-Chromosome%20Variation%20Among%20Sudanese.pdf
http://www.ncbi.nlm.nih.gov/pmc/articles/PMC384897/figure/FG1/ http://en.wikipedia.org/wiki/Haplogroup_J1_(Y-DNA)
http://en.wikipedia.org/wiki/Talk:Haplogroup_J1_(Y-DNA)
Essentially, the issue is whether J1 dominates in Sudan and the Caucasus at over 60%. John Lloyd Scharf 10:18, 23 August 2011 (UTC)

Hi Melcombe. You inserted a OR warning in the article, and I commented on the discussion page, but you did not respond for several months, and so I removed the OR warning. Now you reinserted it. Why not use the discussion page? Do you need further proof? Bo Jacoby (talk) 22:01, 15 September 2011 (UTC).[reply]

Melcombe. While still not using the discussion page you remove useful subsections. That is plain vandalism on your part. Bo Jacoby (talk) 23:44, 20 February 2012 (UTC).[reply]

This is of course nonsense. The material is plainly WP:OR as no WP:Reliable sources have been included. It also fails the test of being important to the description of Bayesian inference in general terms, and serves only to get in the way of expansion in useful directions. Moreover, it fails any reasonable excuse for being included as a "proof" of a result such as outlined as possibilities in WP:MOSMATH. Applying the rules of Wikipedia standards is clearly not vandalism. If this stuff were "useful" someone could provide a source for it and it might then be sensible to construct an article specific to Bayesian inference for that specific distribution, in which this stuff could be included. Melcombe (talk) 22:40, 22 February 2012 (UTC)[reply]

Thank you for answering after 5 months. I will comment on Bayesian inference#Deletion of unnecessary maths portions. Bo Jacoby (talk) 03:51, 23 February 2012 (UTC).[reply]

Hi Melcombe!

Thanks for your help with the article.

(I did everything from memory, so there is some chance of minor mismatches between content in our article and the citation to Lohr or the 3 Swedes, which to me is tolerable.... As you have reminded me before, page references are preferable, but my time is limited ....)

You might wish to look at the template on social surveys, template:Social surveys, which I've tried to upgrade.

Best regards,  Kiefer.Wolfowitz 16:45, 16 September 2011 (UTC)[reply]

I noticed you put the notability tag on spectral risk measure. I have decided over the next few days to try to improve that page to the best of my ability (since I am currently working on my PhD with research on risk measures). I was wondering if you can remember what about that topic made you think it was non-notable. Thank you. Zfeinst (talk) 04:57, 12 October 2011 (UTC)[reply]

Thank you for your comments, I believe I have dealt with them (at a minimum at least) and have removed the notability tag. If you have any other comments I would appreciate it, but you have helped a lot already, thank you.Zfeinst (talk) 17:59, 14 October 2011 (UTC)[reply]

Hi could you revert the changes back for now? It was my intention to change the text to follow "Zero's" copy further down, through the part about definitions, and maybe add my paragraph only as a clarification somewhere further down. I'm pretty new to this, sorry if I'm not putting my comments in the right place. Marc.coram (talk) 10:45, 28 October 2011 (UTC) Oh I see, kinda, you put that material into "other representations". Hmmm. But those are the *real* definitions. Bother, I'm too tired to figure this out right now. I'll return to the MVN chat page tomorrow, and try to draft a fix /discuss it there. Marc.coram (talk) 11:00, 28 October 2011 (UTC) Hi Melcombe. I made some edits last Friday. Seem ok? Thanks for your help. — Preceding unsigned comment added by Marc.coram (talk • contribs) 06:05, 31 October 2011 (UTC)[reply]

Thanks for your recent contributions to the M/G/1 queue article. Particularly in the definition section I feel you've significantly improved readability. I'm less certain about renaming the page though. Google scholar suggests M/G/1 queue being a much more common name for the model. I was tempted to move the M/M/1 model article to M/M/1 queue. Was your rename based just on the existing M/M/1 articles? Gareth Jones (talk) 14:49, 1 November 2011 (UTC)[reply]

Yes, this was based on the existing M/M/1 and M/M/c articles. However, I slightly prefer the M/G/1 queue type of name, so why not go ahead and rename all 3 to that style. Melcombe (talk) 17:53, 1 November 2011 (UTC)[reply]

Also, I've corrected your clarification request about the queue length distribution. Gareth Jones (talk) 17:28, 1 November 2011 (UTC)[reply]

Hi there. You might like to check what is happening to Statistical proof since you've been there before! Tayste (edits) 01:41, 11 November 2011 (UTC)[reply]

Hi. When you recently edited List of statistics articles, you added links pointing to the disambiguation pages Double exponential distribution and Sample standard deviation (check to confirm | fix with Dab solver). Such links are almost always unintended, since a disambiguation page is merely a list of "Did you mean..." article titles. Read the FAQ • Join us at the DPL WikiProject.

It's OK to remove this message. Also, to stop receiving these messages, follow these opt-out instructions. Thanks, DPL bot (talk) 10:53, 8 January 2012 (UTC)[reply]

Dear Sir,

Although indeed all the bugs should be reported to R's developers, I think that these bugs also SHOULD be put in Wikipedia, because a) this is known and reproducible and b) is something of intellectual value. I disagree that Wikipedia should consist of flashy ads for buggy software. Please do not exclude this material. — Preceding unsigned comment added by 75.150.66.10 (talk) 17:46, 10 January 2012 (UTC)[reply]

Hi Melcombe. I see that you have reverted two of my edits, eliminating the use of the template "{{dice}}". I'd like to understand your reasoning for these reverts. I have several concerns with regard to the use of this template: It is confusing to find "{{dice}}" instead of the natural "die" when the singular is wanted, it generates a link, which can result in violations of WP:OVERLINK, it will encourage, it seems to me, the very ignorance that is the root of the problem, and finally I don't see how it, in any way, avoids the problem of editors changing "die" to "dice" inappropriately. Regards, Paul August ☎ 15:28, 11 January 2012 (UTC)[reply]

I see you're improving the EMG distribution page (http://en.wikipedia.org/wiki/Exponentially_modified_Gaussian_distribution). I noticed there is another page of similar content, should it be changed to a redirect? http://en.wikipedia.org/wiki/Exponentially_Modified_Gaussian Shawn@garbett.org (talk) 21:42, 23 January 2012 (UTC)[reply]

A redirect would be good, but it may be worth saving the figure which shows the "chromatography" notation. I notice that you had a comment about characteristic functions having been removed ... I haven't seen where these were, but it would be good to have a formula for the characteristic function. Melcombe (talk) 21:54, 23 January 2012 (UTC)

I think the characteristic function is in one of the earliest references. I can add that. I'll can do the redirect if you don't beat me to it. Shawn@garbett.org (talk) 21:58, 23 January 2012 (UTC)[reply]

I have merged the material from the page I created with this one. The estimators of the parameters should be useful for users of this distribution. DrMicro (talk) 19:18, 24 January 2012 (UTC)[reply]

I had posed a question here. Please ignore what I wrote earlier. I was in error.DrMicro (talk) 23:13, 24 January 2012 (UTC)[reply]

First thank you for your help on the articles I have been working on.

Second a question

You edited the sample variances for the kurtosis and skewness. Normally these statistics are used to test for departures from normality.

for example [[1]]

For that reason are the comments regarding normality required or are they possibly redundant? This is potentially a tricky technical issue and I would value your opinion on it. Thank you in advance. DrMicro (talk) 17:57, 26 January 2012 (UTC)[reply]

Thanks you for your note. This was an issue I was not completely clear on.DrMicro (talk) 20:03, 26 January 2012 (UTC)[reply]

Can you explain why you replaced the 'citation needed' flag in standard deviation? I think it is nonsense to put citations to every easily provable fact that can be checked in many different standard textbooks. What is then, according to you, the wikipedia standard? Why is a citation needed for this mathematical fact and not for all the other mathematical facts, formulae, etc. in this article. Or do you think that those should be tagged also. By the way, what do you think how it looks to the reader? It looks as if there is doubt about these facts, which is absolutely nonsense. KKoolstra (talk) 20:16, 28 January 2012 (UTC)[reply]

Hi, I would like to discuss your recent edit to the Absorbing Markov chain article on its talk page. If you could respond to my question there, that would be great. Thanks :) Bender2k14 (talk) 21:46, 28 January 2012 (UTC)[reply]

Hello again. Do you have another response for me on the talk page? Bender2k14 (talk) 13:00, 19 March 2012 (UTC)[reply]

Hello Melcombe,

Few days ago I place a link at http://en.wikipedia.org/wiki/Odds which was heading to http://www.oddsarchive.com That site is quite unique (provides all sports odds in graphical mode for past and future events) and has no hide advertising or affiliation system included.

I would like to ask you why did not let it there ?

Thank you, Marcel — Preceding unsigned comment added by Psiaii (talk • contribs) 08:37, 1 February 2012 (UTC)[reply]

I see it as an advert for the betting industry in general. It seems out of context in the Odds article and fails WP:EL there. You might try placing the link in Sports betting, which at least has the same context. Melcombe (talk) 09:22, 1 February 2012 (UTC)[reply]

Hello Melcombe,

you attributed to me an opinion which I do not hold. If you ask me whether I agree with all the edits of Brad7777, the answer is probably negative. However, I also do not think they are blatantly inappropriate. There is nothing inherently wrong with changing the category structure, and the word "consensus" is abused in the discussion (or so it seems to me).

I do not retract my statement that you cited, and I think it applies to your comment as well. In this thread, only the comments of Tkuvho and David Eppstein contributed to constructive discussion.

Best regards, Sasha (talk) 22:23, 5 February 2012 (UTC)[reply]

You said (....)

The constructive criticism is that the destructive activity of wholesale arbitrary restructuring of categories should stop. User:Melcombe

Thanks for your reply. I moved the discussion here (to make it easier to follow, please correct me if I have not reproduced your comment precisely).

As to substance, I have nothing more to say.

Best regards,

Sasha (talk) 22:37, 5 February 2012 (UTC)[reply]

Hi, I was you created this cat. How does it distinguish itself from Category:Statistics journals? --Guillaume2303 (talk) 05:19, 10 February 2012 (UTC)[reply]

Melcombe: sorry for stalking (I hope I am not yet non grata here);

Guillaume: although the boundary between probability and statistics is not very sharp, we do keep Category:Probabilistic inequalities and Category:Statistical inequalities, et cet. The journals in the new category (except perhaps for Annals of Mathematical Statistics) are clearly in probability and not in statistics; as to Annals of Mathematical Statistics — it is also reasonable to categorise it in both categories (see its history described in the article). Sasha (talk) 05:39, 10 February 2012 (UTC)[reply]

Just to note that categories have talk pages, just as do articles. But see also list of probability journals. Melcombe (talk) 08:50, 10 February 2012 (UTC)[reply]

• I'm aware they have talk pages, but often people create a cat without putting it on their watchlist, which is why I posted here. I find the difference between statistics and probability a bit difficult to grasp (and me usually priding myself on being statistic-literate - although admittedly in the land of the blind: neuroscientists). To insert "probability journals" in the scientific journals category tree, would you say that they are a subcategory of "statistics journals"? Or better to put them under "mathematics journals"? Or is probability as a field on a par with mathematics? Thanks for further clarification. Also, if I am confused, other editors also might be. To reduce the risk of misclassifications, it would perhaps be helpful to include a short note at the top of the category explaining what exactly it should include. --Guillaume2303 (talk) 09:01, 10 February 2012 (UTC)[reply]

My suggestion would be to leave notes on people's talk pages, and also on project talk pages (here the mathematics and statistics projects at least, although I think there is also a journals project), that a discussion has been started, so that a range of people can join in. See probability and statistics for the distiction between probability and statistic. For the category, I was really thinking of "probability theory" rather than "probability", but retained the name used in the already-existing list of probability journals. I see "probability theory" as entirely contained within "mathematics", in parallel to, but separate from (and partly overlapping "statistics") ... but with "statistics" extending well outside just "mathematics", since it covers things that are not mathematical. Melcombe (talk) 09:21, 10 February 2012 (UTC)[reply]

• You are of course free to start a discussion, but to me this exchnage here is not a "discussion" in the sense of people arguing different sides of an issue. All I'm doing is asking for guidance. "Probability journals" is categorized under "mathematics" journals and "statistical journals" is on the same level as mathematical journals, which seems to square with what you say above. I'll put a link to the "probability and statistics" at the top of the two journal categories, to enlighten other editors :-) --Guillaume2303 (talk) 09:42, 10 February 2012 (UTC)[reply]

Hi, regarding this issue, the problem in general with this kind of lists is that they attract all those crappy OA journals that pop up all over the world, trying to make a fast buck. For an editor not in that particular field, it is very difficult to distinguish what is spam and what is legit. So I routinely remove journals that are either redlinked or (even worse) only have an external link. If you have this list watchlisted and will check for inclusion of non-notable trash, I'm fine with leaving it as it is (and can then safely remove it from my own watchlist, which is full enough), but if this is a one time thing, I'd prefer that journals only are added if they are really notable (i.e. after they have gotten an article). Thanks. --Guillaume2303 (talk) 18:16, 10 February 2012 (UTC)[reply]

The content of articles should be discussed on article talk pages. Melcombe (talk) 11:31, 13 Fe