CODEWITHSUNDEEP
r
Ben
Ben

Reputation: 1522

Sapply function confuses me

It's a basic question and sure, there are a lot of examples in google.. but I just do not understand this small bunch of code..

V <- seq(50, 350, by = 1)

> VK
    Voltage^0     Voltage^1     Voltage^2     Voltage^3 
-1.014021e+01  9.319875e-02 -2.738749e-04  2.923875e-07 

plot(x = V, exp(exp(sapply(0:3, function(x) V^x) %*% VK)), type = "l"); grid()

I tried to get behind this after playing a lot with the function itself but.. I cannot apply my ideas to this certain line. As far as I got I believe I can tell: sapply is a function that applies the body for each element of a vector or list or something similar. In this case this is V. The point that confuses me is the "0:3" part (which seems to be amount of elements of VK) and the end of the function %*% VK. When I do the very same on my own with different numbers than VK is summed up and then used as a coefficient for exp(exp(V^x)). But here in this case this makes no sense. Furthermore: By googling I always read that sapply yields a vector. Due to the fact that the code above generates a plot this is a 2D-vector as result?

Upvotes: 0

Views: 118

Answers (3)

Benjamin
Benjamin

Reputation: 17369

Just for illustration, here is what the code looks like when translated into a for loop.

V <- seq(50, 350, by = 1)

VK <- c("Voltage^0" = -1.014021e+01,
        "Voltage^1" = 9.319875e-02,
        "Voltage^2" = -2.738749e-04,
        "Voltage^3" = 2.923875e-07)

coef <- matrix(0, ncol = 4, nrow = length(V))

for(x in 0:3){
  coef[, x + 1] <- V^x 
}

Y <- exp(exp(coef %*% VK))

plot(V, Y, type = "l"); grid()

It works because sapply is an over achiever that tries not to fail by returning a matrix if it can't return a vector.

Upvotes: 1

Katie
Katie

Reputation: 362

sapply(0:3, function(x) V^x)
> sapply(0:3, function(x) V^x)
       [,1] [,2]   [,3]     [,4]
  [1,]    1   50   2500   125000
  [2,]    1   51   2601   132651
  [3,]    1   52   2704   140608

As stated in the comments, sapply generates a matrix (301x4) where each column represents V^0, V^1,V^2, and V^3. Then, it is multiplied by VK (4x1). This will generate a vector (301x1).

> sapply(0:3, function(x) V^x) %*% VK
               [,1]
  [1,] -6.128411312
  [2,] -6.060636871
  [3,] -5.993320708

After these steps, you apply the exponential function twice to your new vector. This new vector contains your y-values for your plot.

If you wanted to apply "(..sapply(0:2,..)" instead of "(..sapply(0:3,..)", then change VK to this: 

VK <- c(-1.014021e+01,  9.319875e-02, -2.738749e-04)
names(VK) <- c( "Voltage^0",     "Voltage^1",     "Voltage^2")

Upvotes: 1

KenHBS
KenHBS

Reputation: 7164

You are doing multiple things in one line. I guess you are getting lost in the different parts on 1 line where like 5 things happen at the same time. 

plot(x = V, exp(exp(sapply(0:3, function(x) V^x) %*% VK)), type = "l"); grid()

First off, the main part is plot(x, y, type = "l"). This just plots a vector x against a vector y. In your case, y is exp(exp(sapply(0:3, function(x) V^x) %*% VK)).

An easier and prettier way to have written the code, would've been:

inside_of_y <- sapply(0:3, function(x){ V^x %*% VK })
y_variable  <- exp(exp(inside_of_y))

plot(x = V, y = y_variable, type = "l")    
grid()

The first line is probably still messing with you a bit. In essence, sapply is just a for-loop. It is built up of two ingredients: part1 and part2.

part1 is an object with multiple elements, like a data.frame or a vector (here: part1 is equal to c(0, 1, 2, 3)). The elements inside this object are used as arguments for part2. In this particular case, you defined a custom function as part2. In this custom function, you take V to the power of x and multiply the result with the matrix VK. The tricky part is that x is equal to 0 in the first iteration, 1 in the second iteration, ..., and 3 in the last one. You just looped through 0:3, which is part1 of your sapply(part1, part2).

So inside_of_y now contains 4 elements. Then we just proceed with plotting x against y with the plot(x, y) function.

Upvotes: 1

Related Questions