[R-pkgs] short presentation of the new package "GuessCompx"

[Agenis-Nevers Marc]

Dear R community members, please find here a short presentation of our new
package.

We introduce GuessCompx, a new R package that makes an empirical guess on
the time and memory complexities of an algorithm or a function. It will
test multiple, increasing-sizes samples of the user’s data and try to fit
one of seven complexity functions : O(n), O(n2), O(log(n)), etc. Based on a
best fit procedure using LOO-MSE error, it also predicts the full
computation time and memory usage on the whole dataset. It relies on the
base R functions `system.time` and `memory.size `, the latter being only
suitable for Windows users. Together with this result, a plot and a
significance test are also returned.

Complexity is assessed with regard to the user’s actual dataset through its
size (and no other parameter). We provide several examples showing some use
case (distance function, time series, custom function) and how to best tune
the parameters.

The subject of empirical computational complexity has been relatively
little studied in computer sciences, and such a package provides a
convenient and simple procedure to estimate it, thus preventing the user to
run any computation for an unknown amount of time. Empirical fit does not
guarantee to find the true complexity function but approaches it in an
acceptable way. The package does not require to have the code of the target
function.

https://cran.r-project.org/package=GuessCompx

Regards,

Marc Agenis
Neeraj Bokde
marc.agenis using gmail.com
+33 (0)6 19 60 91 23

	[[alternative HTML version deleted]]