lmviz (version 0.2.0)

Simulation.default: Simulates a sample

Description

Simulates an (x,y) sample (suitable for estimating a lm) which can be either non linear/heteroscedastic/non normal or in line with standard lm assumptions

Usage

Simulation.default(model.to.sim)

Arguments

model.to.sim

an integer between 1 and 4 where 1=non linearity; 2=heteroscedasticity; 3=non normality; 4=no violation

Value

A list of objects (of which the first two are essential, the following are needed to display the correct solution in the shiny app)

x,y

the sample

my

the true mean of Y

sderr

the true standard deviation of errors

errore

the errors

xperdens,ferrore

coordinates of points of the true density of errors

Details

The sample size is simulated between 10 and 1000 according to a uniform distribution. The explanatory variable is simulated as uniform, Gaussian, chi.square, t or a mixture of normal distributions. If non linearity or heteroscedasticity is chosen a random regression function or variance function is defined, if non normality is chosen the (always additive) error is simulated from a chi.square, t, Beta or truncated normal.#'

See Also

ComputerDecision.default, checksim

Examples

# NOT RUN {
Simulation.default(1)

Simulation.default(sample(1:4,1))


# }