Data 621 - Blog Post #3

Model Comparison with Tidyverse & Radiant Packages

Libraries

library(tidyverse)
library(dplyr)
library(radiant)
library(kableExtra)

Overview

In my first blog post we discussion, the flexible of the tidyverse package. In this blog post we will be combining the tidyverse and radiant packages for some quick model comparison.

The radiant package is a platform-independent browser-based interface for business analytics in R, based on the Shiny package.

Build Model

Iris Data Set

fit1 <- lm(Sepal.Width ~ Petal.Length + Petal.Width, iris)
fit2 <- lm(mpg ~ wt + qsec, mtcars)

Model Comparison

full_join(glance(fit1), glance(fit2)) %>% 
  add_column(" " = c("Model 1", "Model 2")) %>% 
  select(" ", everything()) %>% 
  kable() %>% 
  kable_styling()

## Joining, by = c("r.squared", "adj.r.squared", "sigma", "statistic", "p.value", "df", "logLik", "AIC", "BIC", "deviance", "df.residual")

	r.squared	adj.r.squared	sigma	statistic	p.value	df	logLik	AIC	BIC	deviance	df.residual
Model 1	0.2131034	0.2023973	0.3892661	19.90490	0	3	-69.80176	147.6035	159.6461	22.27463	147
Model 2	0.8264161	0.8144448	2.5961751	69.03311	0	3	-74.36025	156.7205	162.5834	195.46363	29

The function glance creates a single line data frame from a model best fit summary.