Deviance Deep Dive: Part I
How ClaNC - and `reclanc` - works.
Introduction
In this post, I’m going to give a brief overview into the theory and mechanics of my refresh of the classification package ClaNC, reclanc. The intended audience is those who know just a little about gene expression data and just a little about statistics. If you’re confused, feel free to reach out!
Usage of reclanc won’t be covered here. I’ll be writing a separate vignette - stay tuned.
What is classification, and why do it?
Classification, in essence, requires two steps:
…Peter Vidani's tumblr is the best one even though he's doing it all wrong
Disclaimer
This post has swear words in it. If you don’t like that, don’t read this one. If you don’t like that but you DO like not liking things, dig in boss. If you’re looking to hire me and trying to figure out if I’m a good fit for your company culture, I promise I will step off company premises into the designated swearing area on my swear breaks and get all my swears out then.
…Barely Enough Combinatorics: From 0 to the Binomial Coefficient
Surviving Lawlessness, Pt. 2: Being Discrete
Introduction
This is part 2 of my slow trudge through J. F. Lawlessness’ book “Statistical Models and Methods For Lifetime Data”. Part 1 is here.
Despite our penchant for measuring time in terms of things like ‘seconds’, time is continuous - a smear of unidirectional, infinitesimal, temporally flavored jam across the universe. Sometimes, though, time-to-event things happen in more discrete units. Consider graduation time - students tend to graduate yearly, all at the same time (ie graduation day)1. For this section, we’re going to describe things in terms of graduation, but you might also consider a ‘yearly checkup’ to be another case of discrete-time events.
…Surviving Lawlessness, Pt. 1
Introduction
Like it or not, things are happening all the time. Even worse, you don’t know when a lot of these things will happen. Sure, you might have a general idea - kids learn to walk around the age of 1 or so, students tend to graduate after 4 years of undergraduate school, and dogs learn to shake after a certain number of treats and head scratches - but you don’t know exactly when these things happen. However, if you’re clever, you can work up a handful of tools to help predict when something might happen. You’ve given your dog 20 treats and head pats - what is the probability that he’s finally going to nail the handshake thing on the next one? Or maybe you provide a group of students with a dedicated tutor - do they graduate faster? These are the kinds of questions you can answer if you are willing to wade into the mires of statistics. Don’t worry. We’re going to do this together. I even brought you a snorkel.
…What's a Wilcoxon signed-rank test?
I was tired of looking up what various statistical tests do, so in hopes of remembering, I wanted to try the ’learn by teaching’ method. Worse comes to worst, I can just look back at this if need be.
That was five days ago. After five days of looking at poorly scanned PDFs of books published in the last millennium, formulae that look more like hexes than anything else, diving down into the C source code of R, and, of course, plenty of time on Wikipedia - I think I almost have a grip as to what’s going on.
…