WORK IN PROGRESS ALERT - Using apply functions in R

2 min read

Here is a simple for loop as might be used in a simulation. This is just the first layer and produces one instance of the data. We would run this function a large number of times in our simulation. The function takes some parameters and generates a series of values in each time period. All of data is accumulated in a data.frame which is returned at the end of the function.

# Initialisation of parameters and constants

n.sim <- 1 
n.periods <- 10 
pois.lambda <- 33.4 
pois.se <- 1.8276
gamma.shape  <- 0.4719746  
gamma.shape.se  <- 0.029900 
gamma.rate  <- 0.0097803 
gamma.rate.se  <- 0.000983 
set.seed(3952)

create.dat <- function(n.periods,
                pois.lambda, pois.se,
                gamma.shape, gamma.shape.se,
                gamma.rate, gamma.rate.se){
  
  
  df <- data.frame(period = integer(),
                   y = numeric())

  
  # Create observations

  for(i in 1:n.periods){

    # create n - number of obs in this time period

    n.per <- rpois(1, lambda = rnorm(1, pois.lambda, pois.se))
    
    
    # create y values of the obs this period

    y <- rgamma(n.per, 
                shape = rnorm(1, gamma.shape, gamma.shape.se),
                rate = rnorm(1, gamma.rate, gamma.rate.se))
    
    
    # store period number and x

    df <- rbind(df, data.frame(period = i, x = x))
  }
  
  df
}

An alternative to this is to use lapply. First we construct the function that generates the data for a single period. Then we

create.dat  <- function(x, pois.lambda, pois.se,
                        gamma.shape, gamma.shape.se,
                        gamma.rate, gamma.rate.se){
  
  
  # create n - number of obs in this time period

  n.per <- rpois(1, lambda = rnorm(1, pois.lambda, pois.se))
  
  
  # create x values of the obs this period

  y <- rgamma(n.per, 
              shape = rnorm(1, gamma.shape, gamma.shape.se),
              rate = rnorm(1, gamma.rate, gamma.rate.se))
  
  data.frame(period = x, y = y)
}

# Now run the function of the number of periods.

l1 <- lapply(1:n.periods, create.dat, pois.lambda, pois.se,
       gamma.shape, gamma.shape.se,
       gamma.rate, gamma.rate.se)

df <- do.call("rbind", l1)

apply R

Related