Calculating correlations between response and certain explanatory variables

I want to create a single column that lets me know the correlations for my dependent variable with all of the explanatory variables that I am interested in (all these columns and many more are stored in a data.frame d). By doing cor(d) I can get all the correlations and开发者_高级运维 by doing cor(d$Var1, d$Var2) I can get a single number, but I want to figure out how to get only the Var1 column from the matrix returned by cor(d), with my being able to select the explanatory variables I want included.

---

The cor function can actually do this as well. Suppose we have:

d=data.frame(dependentVar = c(1,2,3),var1=c(-1,-2,-3),var2=c(9,0,5),junk=c(-2,-3,5))

Then this will do the trick:

cor(d[,"dependentVar"], d[,c("var1","var2")])
     var1       var2
[1,]   -1 -0.4435328

It's less efficient (I guess), but you can also do this:

cor(d)["dependentVar", c("var1","var2")]

which computes the full correlation matrix, and then pulls out the subset you want.

---

@DavidR is correct, though R also supports correlation between the columns of X and columns of Y as:

cor(X, Y)

See ?cor for more information.

---

M <- round(cor(College[,-1]),2)

library('corrplot') #package corrplot

We can use circle, square, eclipse also instead of number if we want images instead of direct numbers

corrplot(M, method = "number") #plot matrix

Corelation matrix