telescope.trainrecommender: Trains the recommendation system

In DescartesResearch/telescope: Hybrid Multi-Step-Ahead Forecasting

Description

Trains a recommendations for chossing the best-suited machine learning approach for a given time series. Note that non-seasonal time series are ignored.

Usage

telescope.trainrecommender(timeseries, natural = TRUE, boxcox = TRUE)

Arguments

timeseries	A list of time series
natural	Optional parameter: A flag indicating wheter only natural frequencies (e.g., daily, hourly, ...) or all found frequencies shall be considered.
boxcox	Optional parameter: A flag indicating if the Box-Cox transofrmation should be performed. It is not recommend to disable the transformation. TRUE by default.

Value

An object of the class telescope.recommendation

Examples

telescope.trainrecommender(list(AirPassengers, forecast::taylor))

DescartesResearch/telescope documentation built on Oct. 23, 2021, 9:51 a.m.