In huajiangyy/lm.hua: Multiple Linear Regression

knitr::opts_chunk$set(
  collapse = TRUE,
  comment = "#>"
)

library(lm.hua)
library(broom)
library(bench)

Define a predictor and target vector

Y = matrix(c(1714,1664,1760,1685,1693,1670,1764,1764,1792,1850,1735,1775), ncol=1)
X = matrix(c(2.4,2.52,2.54,2.74,2.83,2.91,3,3,3.01,3.01,3.02,3.07), ncol=1)

To use this function 'linear_regression' to get Y = beta0 + beta1*X + e:

linear_regression(X, Y)

The built-in linear modeling function lm:

lm(Y ~ X)

Comparing beta0 and beta1

beta0_hua = linear_regression(X, Y)[1,1]
beta1_hua = linear_regression(X, Y)[2,1]
beta0_org = tidy(lm(Y ~ X))$estimate[1]
beta1_org = tidy(lm(Y ~ X))$estimate[2]
all.equal(beta0_hua, beta0_org)
all.equal(beta0_hua, beta0_org)
#beta0_org = loc(lm(Y ~ X))

Comparing performance of the functions

bench::mark(linear_regression(X, Y)[1,1], tidy(lm(Y ~ X))$estimate[1], check = function(x,y){all.equal(beta0_hua, beta0_org, tolerance = 1e-2)})
#bench::mark(linear_regression(X, Y), linear_regression(X, Y), check = function(x,y){all.equal(beta0_hua, beta0_org, tolerance = 1e-2)})
#beta0_org = loc(lm(Y ~ X))