ForecastAccuracy.jl

Implementation of forecast accuracy measures.

Installation

To install the package you can use Pkg.add it as follows:

pkg> add ForecastAccuracy

Contributing

Contributions to this package are more than welcome, if you find a bug or have any suggestions for the documentation please post it on the github issue tracker.

When contributing please note that the package follows the JuMP style guide.

Accuracy measures

ForecastAccuracy.maapeFunction
maape(real::Vector{T}, forecast::Vector{T}) where T

Mean Arctangent Absolute Percentage Error. As developed in https://www.sciencedirect.com/science/article/pii/S0169207016000121

source
ForecastAccuracy.maseFunction
mase(real::Vector{T}, forecast::Vector{T}; steps_ahead::Int = 1) where T

Mean Absolute Scaled Error. As developed in https://robjhyndman.com/papers/mase.pdf

source
ForecastAccuracy.mdaseFunction
mdase(real::Vector{T}, forecast::Vector{T}; steps_ahead::Int = 1) where T

Median Absolute Scaled Error. As developed in https://robjhyndman.com/papers/mase.pdf

source
ForecastAccuracy.rmsseFunction
rmsse(real::Vector{T}, forecast::Vector{T}; steps_ahead::Int = 1) where T

Root Mean Squared Scaled Error.

source
ForecastAccuracy.plFunction
pl(real::Vector{T}, quantile_forecast::Vector{T}, tau::T) where T

Pinball loss function defined in terms of the quantile tau

source