Statistics and Statistical Programming (Winter 2017)/R lecture outline: Week 4

From CommunityData
Jump to navigation Jump to search
  • keys stuff for building confidence intervals and p-values:
    • compute a sample standard error just like we did in the book, but in R
    • t.test() with one sample (build a confidence interval)
  • two things I showed in class which are super useful:
    • sort.list()
    • complete.cases()
  • doing something repeatedly:
    • just define a function and then apply it to a list of things
    • if you to output something in the middle: you use the print() function
  • briefly covered:
    • distribution functions: lets focus on *unif(): the key is on page 222 of Verzani
      • The “d” functions return the p.d.f. of the distribution
        • dunif(x=1, min=0, max=3) # 1/3 of the area is the to the left 1
      • The “p” functions return the c.d.f. of the distribution.
        • dunif(q=2, min=0, max=3) #1/(b-a) is 2/3
      • The “q” functions return the quantiles.
        • qunif(p=0.5, min=0, max=3) # half way between 0 and 3
      • The “r” functions return random samples from a distribution.
        • runif(n=1, min=0, max=3) # a random value in [0,3]
  • doing simple simulations with random data
    • runif()
    • rnorm()
  • running quick simulations
    • write a function to repeatedly take the minimum from a sample
    • experiment by changing the size of the sample

Skipped for now[edit]

  • ordered() — really just a type of factor for ordinal data