3inar

library(boot)
library(plyr)

set.seed(22042017) # for reproducibility
generate_data <- function(nsamples=100) {
  x <- rnorm(nsamples,mean=.75, sd=.5)
  y <-  4 + 5*x -3*x^2 + rnorm(length(x), mean=0, sd = 1)
  data.frame(x,y)
}

3inar

library(plyr)

# some very tight high-dimensional data vs some less tight HD data

big <- raply(75, rnorm(50))
small <- raply(75, rnorm(50, sd=1/50))

datas <- rbind(big, small)

colrs <- c(rep("black", 75), rep("red", 75))

3inar

# regression to the mean

n <- 10000
base <- runif(n, 0,10)
dat <- data.frame(x=base+rnorm(n), y=base+rnorm(n))

plot(dat)
abline(0,1, col="grey")

abline(lm(y~x, data=dat))

3inar

library(plyr)


experiment <- raply(10000, function() {
  x <- runif(100, -1, 5)
  y <- 1 + 2* x + rnorm(100,sd = 16)
  
  split_x <- split(x, ceiling(seq_along(x)/5))
  split_y <- split(y, ceiling(seq_along(y)/5))
  

3inar

library(plyr)

set.seed(2017-06-08)

l=0.5

# this distribution has mean and variance both = l above
# importantly it never produces values < 0
plot(table(rpois(100, lambda=l)))

3inar

# This is my experiment that I want to run many times
x <- rnorm(500)
y <- rnorm(500)

my_model <- lm(y~x)

# I usually like to write my results to a randomly named file. I recommend using an absolute path
# so you're 100% sure where the file ends up
random_output_file <-paste0("/home/einar/my_experiments/linear_model", as.character(sample(1:999999, 1)), ".RData")
save(my_model, file = random_output_file)

3inar

#!/usr/bin/env bash

# commands to the batch system start with '#SBATCH '

# pick a name for your job
#SBATCH --job-name=my_job

# I want 500 duplicate jobs that use a single cpu each
#----------------------------------------------------------------
# one machine per job:

3inar

#!/bin/bash
# --- Version history ---
#   0.4: added variable to store file path, and $2 for base file name
#        added variable to store desired reporting interval
#   0.3: added $1 to send in process ID at run time.
#   0.2: switched to $SECONDS for the loop. works.
#   0.1: didn't work well at all.
# --- Version history ---
 
# Usage: cputrack [PID] [filename]