A three-stage approach for modeling multiple time series applied to symbolic quartile data

作者：

Highlights：

• Time series are summarized in classes of entities represented by quartile data.

• More representative classes are obtained based on a prototype selection method.

• Boxplots are built and used for forecasting.

• Synthetic and real data sets and an application confirm the usefulness of the model.

摘要

•Time series are summarized in classes of entities represented by quartile data.•More representative classes are obtained based on a prototype selection method.•Boxplots are built and used for forecasting.•Synthetic and real data sets and an application confirm the usefulness of the model.

论文关键词：Symbolic data analysis,Boxplots,Prototype selection,Mutual information,Multiple time series,Vector autoregressive models

论文评审过程：Received 1 June 2020, Revised 5 September 2021, Accepted 5 September 2021, Available online 25 September 2021, Version of Record 30 September 2021.

论文官网地址：https://doi.org/10.1016/j.eswa.2021.115884