R Linear model predict function for many variables

Question

I have a matrix of variables X and fitted linear model Y ~ X1 + X2 + X3 + X4 . Let's say I want to predict Y for a vector of means of my variables (X1_mean, X2_mean, X3_mean, X4_mean) (and let's call it X_means , I use function colMeans ). So, normally I'd do this in this fashion:

predict(model, X1 = X1_mean, X2 = X2_mean, X3 = X3_mean, X4 = X4_mean)

But is there a faster way to do that? For 4 variables typing (Xi = Xi_mean) is not the end of the world, but for, let's say, 50 variables it will take a while. Can I somehow rewrite it so as to assign each mean to each variable? Something like colnames = means .

Answer 1

Let's say your data looks like this:

set.seed(1)
df <- setNames(as.data.frame(matrix(rnorm(25), ncol=5)), c("Y", paste0("X", 1:4)))
df
#>            Y         X1         X2          X3          X4
#> 1 -0.6264538 -0.8204684  1.5117812 -0.04493361  0.91897737
#> 2  0.1836433  0.4874291  0.3898432 -0.01619026  0.78213630
#> 3 -0.8356286  0.7383247 -0.6212406  0.94383621  0.07456498
#> 4  1.5952808  0.5757814 -2.2146999  0.82122120 -1.98935170
#> 5  0.3295078 -0.3053884  1.1249309  0.59390132  0.61982575

And your model looks like this:

model1 <- lm(Y ~ X1 + X2 + X3 + X4, data = df)

Note that you can get a named list of the means of all your predictor columns like this:

lapply(df[-1], mean)
#> $X1
#> [1] 0.1351357
#> 
#> $X2
#> [1] 0.03812297
#> 
#> $X3
#> [1] 0.459567
#> 
#> $X4
#> [1] 0.08123054

So to predict at the mean of each predictor in this case we could just do:

predict(model1, lapply(df[-1], mean))
#>         1 
#> 0.1292699

^{Created on 2020-12-08 by the reprex package (v0.3.0)}