Hi, I am considering switching from GluonTS to Nix...
# general
c
Hi, I am considering switching from GluonTS to Nixtla to perform probabilistic forecasting of intermittent time series. I need sample paths as output, quantiles are not enough. Which models implemented in Nixtla besides DeepAR can generate sample paths? Thanks!
m
Hello! To my knowledge, DeepAR will be the easiest to generate sample paths. It is possible to use probablistic losses with many models, but then there would be some manual work to generate the sample paths.
👍 1
c
Thank you @Marco