How to find which pairs of predictors are highly correlated in R ? Given are 3 predictors X1, X2, X3 and dependent variable is Y

Question

enter image description here

Tried using vif() function but how to get the correlation between pairs of predictors?

fit<-lm(y ~ X1 + X2 + X3, data=data1) vif(fit)

Answer 1

You can easily find correlations between varibles with the cor function and plot it with the package corrplot

library(corrplot)
df <-
  data.frame(
    x1 = round(runif(100, min = 100, max=170)),
    x2 = sample(1:10, 100, replace = T),
    x3 = round(runif(100), 2),
    y = round(runif(100), 2)
  )

cor(df)
corrplot::corrplot(cor(df),method = "color")

Answer 2

If you just want the correlation coefficients, you can just use cor . If you want them all in a go, use sapply :

sapply(data[, c("X1", "X2")], function(x) cor(x, data$X3))

Answer 3

There are many ways..

Create dummy data

df <-
  data.frame(
    x1 = round(runif(100, min = 100, max=170)),
    x2 = sample(1:10, 100, replace = T),
    x3 = round(runif(100), 2),
    y = round(runif(100), 2)
  )

Using psych package

library(psych)
  pairs.panels(df, 
            method = "pearson", # correlation method
            hist.col = "#00AFBB",
            density = TRUE,  # show density plots
            ellipses = TRUE # show correlation ellipses
            )

Using GGally package

library(GGally)

    ##GGally package
        ## Correlation matirx
    
            ggcorr(df,
            nbreaks = 6,
            label = TRUE,
            label_size = 3,
            color = "grey50")
        ## Correlation Plots
            ggpairs(df, columns = 1: ncol(df), title = NULL,
                upper = list(continuous = "cor"),
                lower = list(continuous = "smooth"),
                mapping = NULL)