I can’t believe how many drafts I’ve been able to keep for more than six months.
% ack "draft\s?[:=] true" content/post/*.md | wc -l 8
Why most positive subgroup analyses are false?
Source: BMJ 2018; 363 doi: https://doi.org/10.1136/bmj.k4245
Back from another trip to my aunt’s house.
Time for Vikings 3 now.
A newcomer in the Clojure land: Machine learning in Clojure with XGBoost.
Thanks to https://archive.org, the CLRS textbook (3rd ed.) can be freely downloaded. One can also find the famous Sedgewick’s Algorithms. I found a PDF copy long ago, and it is dedicated by Simon Plouffe.
There is always something we miss. Here is a shot of those vines that cradled my childhood.
Frank Harrell started hosting his own blog several months ago, and I followed his posts once in a while, that is almost every time I see something new on Twitter. His last annoucement is about a talk he will give at the Johns Hopkins Department of Biostatistics, which appears to be a mix of his latest posts on this topic. I wish I could attend his talk as I did for his RMS workshop in Ottawa some years ago.
Please note that my last tweet or retweet is from October. Don’t expect any other news from me–I will only bookmark tweets that I found interesting while the liking option is still available. If the “like” feature is going to disappear, then it means I will definitely forget about Twitter.
It implements vectorized versions of all C99 real floating point math functions. It can utilize SIMD instructions of modern processors. SLEEF is designed to fully utilize SIMD computation by reducing the use of conditional branches and scatter/gather memory access.
It looks interesting even if I am not going to use it any time soon. (Other than the short vector math library available in clang).
The first season of Vikings is finally over. Let’s start season 2.
Last episode of The Killing 2 planned for tonight.
Here is the Request Map for this website. Still a lot of external dependencies that I could get rid of.
Want to test your competitive programming skills? Take a look at this Bachelor’s Thesis: Analysis and solution of a collection of algorithmic problems (by Rafael Eusebio López Martínez).
In competitive programming, one has to use knowledge in algorithms and data structures to find solutions to algorithmic problems, then put those a ideas into a correct computer program that solves the problem within given time and memory constraints. This activity involves learning about a wide range of complex data structures and algorithms, and many hours of training.
See also my review of the Competitive Programmer’s Handbook.
I don’t know if there’s a better solution, but this is working very well in my
And don’t even look back!
I just came across this new book: Foundations of Data Science (PDF, 479 pp.). It looks great and furthermore this is one of those rare books which provides an original approach to data science and machine learning from rigorous (mathematical) arguments. Other resources are provided in the following Twitter thread.
If you are like me a long time user of command-line tools for data munging, you will probably find some useful utilities on this post: Data Processing Resources: Command-line Interface (CLI) for CSV, TSV, JSON, and XML.
“The numbers 111, 222, 333, 444, 555, 666, 777, 888, and 999 are all evenly divisible by 37, leaving no remainder.” (via @pickover) Using base 10 notation, a three-digit number, say $aaa$, can be written as $100\cdots a + 10\cdot a + a = a \cdot (100 + 10 + 1) = a \cdot 111$. Clearly, 111 is a divisor of $aaa$ ($111 = 37\times 3$), but also of 222, 333, … Hence, all three-digit numbers that are multiple of 111 will fit the bill.
Welcome to season 2 of The Killing! Also, I just had to reboot my Macbook–Mojave upgrade–after 37 days uptime with the same Emacs running in the background.