Following the comments on our earlier submission to PNAS, we have written (and re-arXived) a revised version where we try to spell out (better) the distinction between ABC point (and confidence) estimation and ABC model choice, namely that the problem was at another level for Bayesian model choice (using posterior ... [Read more...]

A link on R-bloggers signaled a series of blogs and videos by IBM Netezza about Thomas Bayes and the consequences of his theorem. Which made me realise this was indeed the 250th anniversary of his death, and that maybe we (as a collective, incl. ISBA) should have done something on ... [Read more...]

The Reverend Thomas Bayes died 250 years ago this month. His grave, located near epidemiological centre of excellence St Mary's College, remains a point of pilgrimage for statisticians (of both Bayesian and Frequentist stripes) visiting London to this day. Because since then, Bayes Theorem has been the underpinning of predictive analytics ... [Read more...]

Introduction Suppose you see two drunks (i.e., two random walks) wandering around. The drunks don’t know each other (they’re independent), so there’s no meaningful relationship between their paths. But suppose instead you have a drunk walking with her dog. This … Continue reading → [Read more...]

If you: use Linux have just upgraded your R installation from 2.12 to 2.13 installed some/all of your packages in your home area (e.g. ~/R/i486-pc-linux-gnu-library/2.12) and… …are wondering why R can’t see them any more just do this: # at a shell prompt cp ~/R/i486-pc-linux-gnu-library/2.12 ~/R/... [Read more...]

This morning I tried to completely update my R installation. I first dumped a list of all the packages I have on my system using the installed.packages() function. Then I installed R 2.13.0 using the OS X disk image. And finally I reinstalled all of my packages from scratch. Unfortunately, ... [Read more...]

For my second lecture today, I need to plot a likelihood surface for a basic two-component mixture with only the means unknown: here is the R code to speed up things llsurf=function(trumyn=2.,wayt=.3,var2=1.,ssiz=500){ # draws the log-likelihood surface and a random sample sd2=sqrt(var2) parti=(runif(... [Read more...]

Look at just the first digit and the number of digits. science: 32914, 11566, 4989, 3743, 968, 814, 673, 482, 286, 2811 black and white: 1694, 1167, 1108, 988, 919, 639, 596, 591, 580, 544 lol: 22627, 18100, 17688, 14374, 13459, 12045, 4711, 3779, 36... [Read more...]

The plyr package (by Hadley Wickham) is one of the few R packages for which I can claim to have used for all of my statistical projects. So whenever a new version of plyr comes out I tend to be excited about it (as was when version 1.2 came out with ... [Read more...]

As I already did two years ago, in connection with the double degree between UAM and Dauphine, I will give a short graduate course at the Universidad Autonoma de Madrid (UAM). It will be part of the regular fourth year statistics course and will focus on mixtures, as given in ... [Read more...]

Once in a while, you embark on what looks like a simple computational procedure only to encounter frustration very early on. “I can’t even read my file into R!” you cry. Step back, take a deep breath and take note of what the software is trying to tell you. ... [Read more...]

Following my earlier post on Buffon’s needle and Bertrand’s paradox, above are four outcomes corresponding to four different generations (among many) of the needle locations. The upper right-hand side makes a difference in the number of hits (out of 1000). The R code corresponding to this generation was made ... [Read more...]

I recently received an email about forecasting and rules of thumb. "Dans la profession [...] se transmet une règle empirique qui voudrait que l'on prenne un historique du double de l'horizon de prévision : 20 ans de données pour une prévision [Read more...]

When Peter Diggle gave his “short history” of spatial statistics this morning (I typed this in the taxi from Charles de Gaulle airport, after waiting one hour for my bag!), he started with a nice slide about Buffon’s needle (and Buffon’s portrait), since Julian Besag was often prone ... [Read more...]

Are MLB players better hitters now than they were 20 years ago? Revolution Analytics' Joseph Rickert uses R to take a look at the data, and offers an instructive lesson in checking your assumptions for statistical tests in the process -- Ed. Data are everywhere – but, even for simple things, I ... [Read more...]

Another mathematical puzzle from Le Monde that relates to a broken calculator (skipping the useless tale): Given a pair of arbitrary positive integers (x,y) a calculator can either substract the same integer [lesser than min(x,y)] from both x and y or multiply either x or y by 2. ... [Read more...]

(This post got published on The Statistics Forum yesterday.) The short book review section of the International Statistical Review sent me Raquel Prado’s and Mike West’s book, Time Series (Modeling, Computation, and Inference) to review. The current post is not about this specific book, but rather on why ... [Read more...]

In what seems like an endless cuRse, I found this week I had to re-grade a dozen R exams a TA’s did not grade properly! The grades I (X) got are plotted below against those of my TA (Y). There is little connection between both gradings… As if this ... [Read more...]

The mathematical puzzle from the weekend edition of Le Monde from a few weeks ago was not too hard to solve by induction but my R code failed miserably! The puzzle was as follows: A calculator is broken in such a way that it starts by exhibiting 0, then pressing 4, 6 or 0 ... [Read more...]

[UPDATE 3/28/2011: Fixed an enormous bug in the R code.] I’m trying to collect data sets that showcase how the classical statistical distributions appear in modern contexts. I’ve already got some data that shows how the gamma distribution appears in video game scores, and now I’m hoping to ... [Read more...]