Pre-SMATS: A multi-task learning based prediction model for small multi-stage seasonal time series

作者：

Highlights：

• The Pre-SMATS model is proposed to predict small multi-stage seasonal time series.

• The model consists of a feature extractor, a classifier and a predictor.

• We propose a new approach for processing seasonal characteristics.

• Pre-SMATS can reduce the impact of insufficient data by enhancing data features.

摘要

•The Pre-SMATS model is proposed to predict small multi-stage seasonal time series.•The model consists of a feature extractor, a classifier and a predictor.•We propose a new approach for processing seasonal characteristics.•Pre-SMATS can reduce the impact of insufficient data by enhancing data features.

论文关键词：Small seasonal time series,Multi-task learning,Data prediction,Neural networks,Multi-stage,Feature enhancement

论文评审过程：Received 22 July 2021, Revised 13 February 2022, Accepted 28 March 2022, Available online 12 April 2022, Version of Record 22 April 2022.

论文官网地址：https://doi.org/10.1016/j.eswa.2022.117121