Cool Statistical Term Names


Less is more. Stay pure. Stay poor.
I thought it might be fun to draft a list of cool names for statistical terms/procedures. Since some of us may not be familiar with them or know their meanings.

Bootstrap (a play on pulling yourself up by your bootstrap - resampling procedure)
Random Forrest (using bootstrap while performing regression trees)
BAgging (bootstrap aggregating with regression trees, bootstrapping variables as well)

I can't remember whether the Jackknife has a cool definition or not?


Dark Knight
BLUE (best linear unbiased estimator)
MIDAS (Mixed data sampling)
Hat Matrix
Agnes (agglomerative nesting : clustering)
Diana (DIvisive ANAlysis :Clustering)

(will add the details later )


Cookie Scientist
Hat matrix would be an awesome Halloween costume. You could carry around a bunch of cheap hats and put them on people at the Halloween party.

How about: Deep Learning


Less is more. Stay pure. Stay poor.
Or what about a matrix filled with different types of hats.

Hat_matrix <- (baseball_hat, cowboy_hat, beret, ski_mask, rows=2)


Fortran must die
You know you made it in academics when you can coin acronyms :p

HVRPFSUDAPCR is not a stats term but is a military acronym I believe...if you know what it stands for (its a type of ammo) you seriously need to get a life :p


Less is more. Stay pure. Stay poor.
In causal diagrams

Collider/M-Bias (inverted fork):
variable with respect to a particular path and there are two variables along the path with arrows that both point to the it (collider).

Confounder: common ancestor of two connected variables

Butterfly Bias: causal model with both confounding and M-bias (looks like a butterfly)


Dark Knight
t-test (tea test)
VARMA (vector ARMA) - An Indian name
NOMINTAE (Nominal Three-Step Estimation)
ARCH Model
MARCH (Multiplicative ARCH)
STAR (Smooth Transition Autoregressive)
Kitchen sink
zero inflated poisson