florianhartig

# The purpose of this script is to show that RF variable importance
# will split importance values for collinear variables evenly,
# even if collinearity is low enough so that variables are sepearable 
# and would be correctly separameted by an lm / ANOVA 

set.seed(123)

# simulation parameters
n = 3000
col = 0.7

florianhartig

library(gdata)
Data = read.xls("http://www.pnas.org/content/suppl/2014/05/30/1402786111.DCSupplemental/pnas.1402786111.sd01.xlsx", 
                nrows = 92, as.is = TRUE)

library(glmmTMB)
originalModelGAM = glmmTMB(alldeaths ~ scale(MasFem) * 
                          (scale(Minpressure_Updated.2014) + scale(NDAM)), 
                           data = Data, family = nbinom2)

# Residual checks with DHARMa

florianhartig

library(mvtnorm)   # to draw multivariate normal outcomes
library(R2jags)    # JAGS-R interface

# function that makes distance matrix for a side*side 2D array  
dist.matrix <- function(side)
{
  row.coords <- rep(1:side, times=side)
  col.coords <- rep(1:side, each=side)
  row.col <- data.frame(row.coords, col.coords)
  D <- dist(row.col, method="euclidean", diag=TRUE, upper=TRUE)

florianhartig

n = 50
x = seq(-1,1, len = n)
f = function(x) 0.3 * x^2
y = f(x) + rnorm(n, sd = 0.2)


par(mfrow = c(1,2))


lmFit <- lm(y ~ x + I(x^2))

florianhartig

trueA = 5
trueB = 0
trueSd = 10
sampleSize = 31

# create independent x-values
x =(-(sampleSize-1)/2):((sampleSize-1)/2)
# create dependent values according to ax + b + N(0,sd)
y = trueA * x + trueB + rnorm(n=sampleSize,mean=0,sd=trueSd)

florianhartig

# modified from http://stackoverflow.com/questions/1401904/painless-way-to-install-a-new-version-of-r-on-windows

# run on old computer / r version
setwd("/Users/Florian/Dropbox/temp") # or any other existing temp directory
packages <- installed.packages()[,"Package"]
save(packages, file="Rpackages")


# run on new computer / r version
setwd("/Users/Florian/Dropbox/temp") # or any other existing temp directory