MSMAPE: MSMAPE - Multivariate Symmetric Mean Absolute Percentage...

In jdestefani/MEMTS: Multivariate Error Measures for Time Series forecasting

Description

Computes the multivariate SMAPE based on the individual time series (according to https://arxiv.org/pdf/1905.03806.pdf)

Usage

MSMAPE(X, X_hat)

Arguments

X	Target matrix (hxk) for k variables and forecasting horizon h
X_hat	Predicted matrix (hxk) for k variables and forecasting horizon h

Value

List containing:

• values_vec: Vector of the individual SMAPEs

• mean: Mean of the vector of the individual SMAPEs

• article_ref: SMAPE computed according to the reference article

Examples

X <- matrix(rnorm(100),10,10)
X_hat <- matrix(rnorm(100,mean=1),10,10)
results_list <- MSMAPE(X,X_hat)
results_list$values_vec # For individual WAPE
results_list$mean # For the mean of the individual WAPE
results_list$mean # For the value computed according to the reference article

jdestefani/MEMTS documentation built on Dec. 20, 2021, 10:04 p.m.