createDataExt: Simulate test data
In rbslandau/ecoteach: Simulation of ecological data for teaching

Description Usage Arguments Note Author(s) Examples

View source: R/createData.R

This function creates synthetic dataset with various problems such as overdispersion, zero-inflation, etc. and was modified to use also externally generated environmental data.

createDataExt(replicates = 1, sampleSize = 10, extPredictors = NULL,
  intercept = 0, fixedEffects = 1, quadraticFixedEffects = NULL,
  numGroups = 10, randomEffectVariance = 1, overdispersion = 0,
  family = gaussian(), scale = 1, cor = 0, roundPoissonVariance = NULL,
  pZeroInflation = 0, binomialTrials = 1, temporalAutocorrelation = 0,
  spatialAutocorrelation = 0, factorResponse = F)

`replicates`	number of datasets to create
`sampleSize`	sample size of the dataset
`extPredictors`	dataframe of environmental data generated externally
`intercept`	intercept (linear scale)
`fixedEffects`	vector of fixed effects (linear scale)
`quadraticFixedEffects`	vector of quadratic fixed effects (linear scale)
`numGroups`	number of groups for the random effect
`randomEffectVariance`	variance of the random effect (intercept)
`overdispersion`	if this is a numeric value, it will be used as the sd of a random normal variate that is added to the linear predictor. Alternatively, a random function can be provided that takes as input the linear predictor.
`family`	family
`scale`	scale if the distribution has a scale (e.g. sd for the Gaussian)
`cor`	correlation between predictors
`roundPoissonVariance`	if set, this creates a uniform noise on the possion response. The aim of this is to create heteroscedasticity
`pZeroInflation`	probability to set any data point to zero
`binomialTrials`	Number of trials for the binomial. Only active if family == binomial
`temporalAutocorrelation`	strength of temporalAutocorrelation
`spatialAutocorrelation`	strength of spatial Autocorrelation
`factorResponse`	should the response be transformed to a factor (intended to be used for 0/1 data)

The basic structure of this function was taken from createData in the package DHARMa by Florian Hartig.

Lisa Huelsmann

require(lme4)


# complex environment

testData = createDataExt(sampleSize = 2000, intercept = 5, fixedEffects = c(2, 2, 0), cor = 0.5)

pairs(testData[, grepl("Environment", names(testData))])

fittedModel <- lmer(observedResponse ~ Environment1 + Environment2 + Environment3 + (1|group), data = testData)
summary(fittedModel)



# External predictors

nvar <- 4
mu <- sample(seq(0.01, 0.5, length.out = 50), nvar, replace = TRUE)
data_sim <- corrEnv(n = 2000, nvar = nvar, ngrad = 3, mu = mu, rho = 0.9, rho.non.corr = 0)

testData = createDataExt(sampleSize = 2000, extPredictors = data_sim[[2]], fixedEffects = c(2, 2, 1, 0))

hist(testData$observedResponse)
pairs(testData[, grepl("Environment", names(testData))])

fittedModel <- lmer(observedResponse ~ Environment1 + Environment2 + Environment3 + Environment4 + (1|group), data = testData)
summary(fittedModel)



### FROM DHARMa



# with zero-inflation

testData = createDataExt(sampleSize = 500, intercept = 2, fixedEffects = c(1),
                      overdispersion = 0, family = poisson(), quadraticFixedEffects = c(-3),
                      randomEffectVariance = 0, pZeroInflation = 0.6)

par(mfrow = c(1,2))
plot(testData$Environment1, testData$observedResponse)
hist(testData$observedResponse)


# binomial with multiple trials

testData = createDataExt(sampleSize = 40, intercept = 2, fixedEffects = c(1),
                      overdispersion = 0, family = binomial(), quadraticFixedEffects = c(-3),
                      randomEffectVariance = 0, binomialTrials = 20)

plot(observedResponse1 / observedResponse0 ~ Environment1, data = testData, ylab = "Proportion 1")


# spatial / temporal correlation

testData = createDataExt(sampleSize = 100, family = poisson(), spatialAutocorrelation = 3,
                      temporalAutocorrelation = 3)

par(mfrow = c(1,2))
plot(log(observedResponse) ~ time, data = testData)
plot(log(observedResponse) ~ x, data = testData)