Gelman on Meehl

Published by on March 23, 2015

Today, Andrew Gelman posted about a paper written over thirty years ago by Paul Meehl. Even though I believe actuaries are less exposed than social science statisticians to some of the pitfalls of what Gelman calls the “garden of forking paths,” we are still exposed, and I’d recommend reading Gelman’s posts on the issue. I Read the full article…

I’ve updated the directions as to how to build 64-bit R for Windows linking to OpenBLAS, as the entire R build process for Windows will change for R version 3.2 and subsequent. I’ve turned it into a static page, instead of a post, in case the build process changes again. While the title has 3.2 in it, the URL doesn’t, so it should be future-proof. Then again, that’s number 47 on the list of 996 famous last words! As always, I’m interested in hearing if anyone successfully builds it and how much of a difference it makes.

Flavors of risk: process, parameter, model, and unknown

Published by on September 15, 2014 | 1 Response

I’ve recently been asked a number of times as to the differences between various sources of risks one is exposed to when creating a model, in specific, for the difference between parameter and model risk. I’ve used the following allegory a number of times to explain how I see the differences between process, parameter, model, Read the full article…

Big Data, Correlations, and the Hales-Jewett Theorem

Published by on May 21, 2014 | 3 Responses

“Big Data” has become one of the most heralded tools, or at least buzzwords, of this decade. The use of “big data” is trumpeted by many studies in almost all disciplines, and, too an extent, understandably so. Hand-in-hand with the geometric progression of computing power has been the development of algorithms that can parse huge Read the full article…

The Heartbleed vulnerability, whose scope may safely be described as epic, should come as reminder to all of us to adopt better password handling and usage. It is well worth reading these password-related essays by two of the better-known cybersecurity experts: Bruce Schneier and Brian Krebs. Personally, I use KeePass2 and the KeeFox plug-in for FireFox, so I can have 20-digit long passwords from a 94-character space and not have to remember any of them. It would be prudent for everyone to change their passwords, at least the important ones (financial, e-mail, etc.), and seriously consider some kind of password manager so that we don’t have to remember the zillion-odd passwords we need.

I’ve re-written the previous post a bit and now it is posted in the Rcpp gallery. I’ve had some really good comments, so I’ll probably have to post an updated version. Keep the comments coming!

Condorcet ranking and Rcpp

Published by on March 13, 2014 | 9 Responses

I had the opportunity recently to be asked to join a committee which had to select candidates for an upcoming election. In the candidate selection phase, the members of the committee use preferential voting to rank the candidates. Prior to the meeting, I did some research on preferential voting systems. I found two main classes Read the full article…

Published again!…sort of

Published by on February 18, 2014

Recently, I found I was published again…sort of. About two years ago, I wrote a review of Gray & Pitts’s Risk Modelling in General Insurance: From Principles to Practice for the South African Actuarial Journal, which was finally published in the 2013 issue. The review itself can be found here (PDF). Enjoy!

R programming tips

Published by on January 15, 2014

After reinventing the wheel a few times, I decided it would be a good idea to keep a list of programming tips or tricks that I find useful, so I started a static page here to keep these tips. It will begin rather haphazardly, but if it grows, I will probably categorize and re-order it. Read the full article…