EVEN MORE FINAL EDIT: You can find the transcript here.

FINAL EDIT: All the details have been settled, so here is a brief summary.

The second CVJC will take place on 1.21.2011 (UTC -- this will start on 1.20 in some tzs).

We will be discussing the paper suggested by Gavin Simpson, "A Statistical Analysis of Multiple Temperature Proxies: Are Reconstructions of Surface Temperatures Over the Last 1000 Years Reliable?" by B. B. McShane and A. J. Wyner and general statistical view on analyzing climate.

The formula is such that the whole day (0:00UTC--23:59UTC) will be devoted to conversations about (and around) it, yet there will be three "concentration" sessions to seed the discussion -- each participant is expected to appear at at least one of them:

  • A at 1:00UTC
  • B at 9:00UTC
  • C at 16:00UTC

If you are planning to come, please register on this page; you'll get an e-mail remainder and others would know who to expect and when.

ORIGINAL QUESTION: There was not too crowdy on the first CVJC, so it is time to fix it with the second one ;-)

To remind for those who are new to the idea:

  • CVJC is a hour-long whole day meeting on chat where we discuss some paper and its theoretical/practical surroundings.
  • The paper must be OpenAccess or a (p)reprint suggested previously on a meta thread like this one and selected in voting.
  • We (mods) would try to invite the author (it didn't work the first time, but we will keep trying).

So, please suggest papers (each in one answer) and this time also time/date. Assuming about 2 week regime I would suggest somewhere between holidays and new year; and please take timezones into account. The deadline for the paper suggestions is 23:59UTC 12.26.2010.

EDIT: Let's say it has been officially moved to the beginning of 2011. I'm going to arrange some precise time/date voting after the paper selection will be done, but please keep suggestions coming.

EDIT2: Thanks to Shane, the new plan is to make CVJC whole day event full of smaller or longer discussions about the paper. Yet, there will be 3 "concentration sessions", at 1:00UTC, 9:00UTC and 16:00UTC (each is selected so that people from 2 out of 3 main timezones may participate), to make people actually meet and start talking.

EDIT3: There was a tie in votes, so I have performed a fair coin roll and selected the climate one.

share

3 Answers 3

up vote 15 down vote accepted

Something topical, controversial and statistical (and unashamedly from [close to] my field) is the Hockey-stick reconstruction of climate over the past 1500 years. A new paper by McShane and Wyner to be published in Annals of Applied Statistics has generated a lot of discussion and has been made a discussion paper, with 13 responses. All can be found here and currently open-access (Scroll to bottom of page):

The main paper and some of the discussions give R code to run the analyses. topic involves time series analysis and multivariate calibration. One potential downside, the main paper is somewhat lengthy and if we add in the discussion papers...!?, plus (at least to my non-statistician non-Bayesian brain) some parts of the main paper and discussion points are quite technical from the stats point of view.

These works represent the first real coming together of statisticians and palaeoecologists to tackle a question of fundamental importance; does the climate of present differ substantially from that experienced in the recent past.

As for dates, I'd favour pushing the next session into early Jan (7th) as I might be hard pushed to get some free time between Christmas and New Year to contribute to whatever topic is chosen, but I am free on Jan 7th to contribute.

share
    
+1 Thanks for the pointer. I had missed this. Looks very interesting. –  fabians Dec 22 '10 at 14:56
3  
To get started, I recommend skimming the paper (pay attention to Figure 14) and then jumping directly to the rejoinder to see where the controversy is. Use the remarks in the rejoinder to go back to any of the discussion papers that look interesting. –  whuber Dec 23 '10 at 17:35
1  
+1, both topical and controversial; see also Hockey stick controversy –  denis Dec 25 '10 at 17:41

What about the following article on collaborative filtering (from what I seem to remember, there were some questions dealing with that topic in the past)?

Xiaoyuan Su and Taghi M. Khoshgoftaar, A Survey of Collaborative Filtering Techniques, Advances in Artificial Intelligence, Volume 2009 (2009).

The recommenderlab R package, from Michael Hahsler, seems to implement some algorithms for recommendation systems (see also the accompanying vignette).

About a possible date, I would suggest avoiding Friday the 31st. So, if the next CVJC has to be held next week, I would rather suggest Wednesday or Thursday. But maybe we could expect a better participation rate the week after; that is, Friday the 7th?

share
    
Looks nice. Due to the date, 31st is of course out of question, but there is no constraint that it must be on Friday. –  mbq Dec 21 '10 at 0:21
    
sounds good –  suncoolsu Dec 21 '10 at 8:31
6  
Yes, I rather the idea of pushing this into January... –  Shane Dec 21 '10 at 18:29

One thought would be a paper that does a statistical analysis of the holidays in some way, such as spending habits or New Year's resolutions. I don't have any great ideas here. One example is:

share

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .