In haisx/basiclm: Linear_model

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

# Install package from GitHub
# devtools::install_github("haisx/basiclm")

# Load the package
library(basiclm)

There is one linear_model() function in the package. It presents some important statistics of linear regression model. The input variables should all be numeric variables.

The Parameters of Function `linear_model`

formula: The formula of interests. It should follow the format of y~x. You can add more than one explanatory variables with different form like y~x+z+I(z^2). e.g.. linear_model(y~x)

data: The name of the data needed to conduct the linear regression model. e.g. linear_model(y~x, data = mydata)

digit: The digits you want to keep in the output. The p-value would not be affected by digit. The default digit option is digit = 3. e.g. linear_model(y~x, data = mydata, digit = 8)

detailed: The detailed means a detailed output of the linear regression model. The detailed table includes more information about coefficients and model like Std Error, t value of the coefficients, R-Squared of the model, and so on. The default detail = FALSE. e.g. linear_model(y~x, data = mydata, digit = 8, detailed = TRUE)

Use `linear_model()` with Vectors

We can create our own response and explanatory variables and put them into the function.

y = c(1:10)
x = c(2:5, 11:6)

linear_model(y~x)

we can get what we called in function and the coefficients of the linear regression model directly.

Use `linear_model()` with Dataset

Here we use iris data as a simple example.

head(iris)

# We will use the numeric variables in iris dataset
linear_model(Petal.Length~Sepal.Width ,data = iris)

# Keep more digits of results
linear_model(Petal.Length~Sepal.Width ,data = iris, digit = 8)

Get Specific Output by Using `detailed` Option

linear_model(Petal.Length~Sepal.Width ,data=iris, detailed = TRUE)

Performance of `linear_model()` vs. `lm()`

set.seed(222)
testx = runif(5000000, 1, 10000)
testy = runif(5000000, 1, 10000)

system.time(lm(testy~testx))
system.time(linear_model(testy~testx))


set.seed(223)
testx = runif(5000000, 1, 10000)
testy = runif(5000000, 1, 10000)
testz = runif(5000000, 1, 10000)

system.time(lm(testz~ testx + testy))
system.time(linear_model(testz~ testx + testy))

Output Comparisons `linear_model()` vs. `lm()`

#iris sample data
linear_model(Petal.Length~Sepal.Width, data=iris, digit = 8, detailed = T)

summary(lm(Petal.Length~Sepal.Width, data=iris))

# Other random generated data
set.seed(2212)
testx = c(1:1000)
testy = runif(1000, 1, 10000)

linear_model(testy~testx, digit = 5, detailed = T)

summary(lm(testy~testx))