In pderdeyn/Stats230pieter: What the Package Does (One Line, Title Case)

knitr::opts_chunk$set(echo = TRUE)

Multiply Some Matrices

Let's multiply matrices. We use a function, mat_mult(), from my package available at https://github.com/pderdeyn/Stats230pieter. mat_mult computes the product $A\times B\times x$. This function uses an optional boolean argument, left, which determines whether the left product, $A\times B$, or the right product, $B \times x$, is computed first. The default argument is left=TRUE.

#install.packages("../",repos=NULL,type="source")
#devtools::install_github('https://github.com/pderdeyn/Stats230pieter')
library(Stats230pieter)
values<-seq(0,100)
A<-matrix(sample(values,100),ncol=10)
B<-matrix(sample(values,100),ncol=10)
x<-c(sample(values,10))
mat_mult(A,B,x)
mat_mult(A,B,x,left=FALSE)

Benchmarking

Let's benchmark multiplying matrices

library(bench)
b<-bench::mark(mat_mult(A,B,x,left=TRUE),
            mat_mult(A,B,x,left=FALSE))
plot(b)

Hence, it appears from this experiment that computing the right product, $B\times x$, first leads to faster computation. It is interesting to see that this expression has multimodal distribution, where the left product variation has just one peak.

pderdeyn/Stats230pieter documentation built on March 21, 2022, 6:32 a.m.