In liulim-Liu/lin.Reg: Univaraite or Multivariate Linear Regression

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

library(lin.Reg)

Decription of `lin.Reg`

This is the function will you can perform univariate or multivariate linear regression models. A second function of summary.lin.Reg is also included for displaying the results of linear regression.

Examples

using for univariate

set.seed(2021)
X = rnorm(100)
Y = X * 3 + rnorm(5)

model_uni = lin.Reg(Y,X)

Using summary.lin.Reg to display the result:

summary_lin.Reg(model_uni)

using for multivariates

set.seed(2021)
X = matrix(rnorm(100), 50, 2)
Y = matrix((X[,1]+ 5*X[,2] + rnorm(50,sd=2)), 50, 1)

model_multi = lin.Reg(Y,X)

Using summary.lin.Reg to display the result:

summary_lin.Reg(model_multi)

Troubleshoot

In same cases, the function will not perform correctly, several error message will then show up

For example, if the response variable are provided more then 1 columns, which is not allowed in this model, a error message will show like this:

set.seed(2021)
X = matrix(rnorm(50), 5, 2)
Y = matrix((X[,1]+ 5*X[,2] + rnorm(50,sd=2)), 100, 2)

try(lin.Reg(Y,X))

Benchmark

comparing with `lm()`

X1 = rnorm(100)
X2 = rnorm(100)
X= cbind(X1,X2)

Y = X1+ 5*X2 + rnorm(100,sd=5)

#if (!require("bench")) install.packages("bench")
benchmark <- bench::mark(lm(Y~X1+X2)$coefficients, 
                     lin.Reg(Y,X)$coefficients)
benchmark